Centimeter-Scale 2D van der Waals Vertical HeterostructuresIntegrated on Deformable Substrates Enabled by Gold SacrificialLayer-Assisted GrowthMd Ashraful Islam,†,‡ Jung Han Kim,§ Anthony Schropp,† Hirokjyoti Kalita,†,‡ Nitin Choudhary,†

Dylan Weitzman,† Saiful I. Khondaker,†,‡,∥ Kyu Hwan Oh,§ Tania Roy,†,‡,∥,⊥ Hee-Suk Chung,*,#

and Yeonwoong Jung*,†,‡,⊥

†NanoScience Technology Center, ‡Department of Electrical and Computer Engineering, ∥Department of Physics, ⊥Department ofMaterials Science and Engineering, University of Central Florida, Orlando, Florida 32826, United States§Department of Materials Science and Engineering, Seoul National University, Seoul 08826, South Korea#Analytical Research Division, Korea Basic Science Institute, Jeonju 54907, Jeollabuk-do, South Korea

*S Supporting Information

ABSTRACT: Two-dimensional (2D) transition metal dichalcogenides(TMDs) such as molybdenum or tungsten disulfides (MoS2 or WS2)exhibit extremely large in-plane strain limits and unusual optical/electrical properties, offering unprecedented opportunities for flexibleelectronics/optoelectronics in new form factors. In order for them to betechnologically viable building-blocks for such emerging technologies, itis critically demanded to grow/integrate them onto flexible or arbitrary-shaped substrates on a large wafer-scale compatible with the prevailingmicroelectronics processes. However, conventional approaches toassemble them on such unconventional substrates via mechanicalexfoliations or coevaporation chemical growths have been limited tosmall-area transfers of 2D TMD layers with uncontrolled spatial homogeneity. Moreover, additional processes involving aprolonged exposure to strong chemical etchants have been required for the separation of as-grown 2D layers, which isdetrimental to their material properties. Herein, we report a viable strategy to universally combine the centimeter-scale growth ofvarious 2D TMD layers and their direct assemblies on mechanically deformable substrates. By exploring the water-assisteddebonding of gold (Au) interfaced with silicon dioxide (SiO2), we demonstrate the direct growth, transfer, and integration of 2DTMD layers and heterostructures such as 2D MoS2 and 2D MoS2/WS2 vertical stacks on centimeter-scale plastic and metal foilsubstrates. We identify the dual function of the Au layer as a growth substrate as well as a sacrificial layer which facilitates 2Dlayer transfer. Furthermore, we demonstrate the versatility of this integration approach by fabricating centimeter-scale 2D MoS2/single walled carbon nanotube (SWNT) vertical heterojunctions which exhibit current rectification and photoresponse. Thisstudy opens a pathway to explore large-scale 2D TMD van der Waals layers as device building blocks for emerging mechanicallydeformable electronics/optoelectronics.

KEYWORDS: 2D materials, TMD, MoS2/WS2, van der Waals heterostructure, layer transfer, flexible device

Two-dimensional (2D) transition metal dichalcogenides(TMDs) such as molybdenum (or tungsten) disulfides

(MoS2 or WS2) offer a rich set of extraordinary materialproperties benefiting from their unique anisotropic structureand near atom thickness. Among them, a combination of largeplanar elasticity [> ×4 in-plane strain limit over silicon (Si)]and tunable band gap energies (∼1.2−1.8 eV) makes themparticularly promising for flexible/stretchable electronics andoptoelectronics.1−4 2D heterostructure layers composed ofvertically stacked dissimilar 2D TMDs held via weak van derWaals (vdW) attractions offer unique 2D/2D interfaces that areenvisioned to exhibit exotic properties unattainable in theirmonocomponent counterparts.5−10 To broaden their scientific/technological versatilities for scaled-up flexible/stretchabletechnologies, a few challenges associated with materials

preparation/integration need to be overcome: (1) They shouldbe grown on a large wafer scale with controlled layerchemistries/morphologies, and the growth methods should beintrinsically scalable. (2) They should present mechanicalflexibility either by being directly grown on flexible substrates orbeing transferred onto secondary flexible substrates. (3) Iftransferred, they should maintain the structural/chemicalintegrities of their original as-grown states. Accordingly, it iscritically demanded to develop a reliable strategy to universallysatisfy all the aforementioned growth/transfer requirements on

Received: June 30, 2017Revised: September 22, 2017Published: September 25, 2017

Letter

pubs.acs.org/NanoLett

© 2017 American Chemical Society 6157 DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

a wafer scale for the current microelectronics manufacturingprocesses. The conventional approaches to manually exfoliate/stack up individual 2D TMD layers suffer from very small areasof transferred layers and thus remain nonscalable.11−22

Chemical vapor deposition (CVD) growths based on thecoevaporation/reaction of metal- and/or chalcogen-containingprecursors generally produce materials with limited spatialcoverages and nonuniform morphologies.23−29 Besides unsat-isfactory growths, a reliable lift-off/transfer of the as-grown 2DvdW heterostructure layers from their growth substrates(mostly SiO2) to secondary substrates is another majorchallenge. The separation of 2D TMD heterostructure layersfrom their SiO2/Si growth substrates has generally involved thedirect chemical etching of the SiO2 using acid/base agents suchas buffered oxide etchant (BOE) or potassium hydroxide(KOH).30,31 This chemical agent-based etching becomesparticularly problematic when the size of the 2D TMD layersto be transferred increases as a large amount of time isdemanded for the complete etching of SiO2 by interpenetrating2D TMD/SiO2 interfaces. For example, ∼6 h of SiO2 etchinghas been reported for CVD-grown vertically stacked MoS2/WSe2 of >2 × 2 cm2.30 The prolonged exposure of 2D TMDsto the strong chemical agents is known to deteriorate their

material properties, motivating the exploration of alternativemethods.32−37 Besides the growth/transfer challenges, it ishighly desirable to directly integrate large-scale 2D layers onmechanically flexible dielectric materials (mostly thin-filmoxides), as demanded in a variety of scaled-up flexible electronicdevices.In this Letter, we report a viable strategy to directly integrate

centimeter-scale 2D TMD vdW heterostructures on “trans-ferable substrates” in stacks of SiO2/Au layers and their faciletransfer to arbitrary substrates without using strong chemicalagents. Motivated by the enhanced debonding nature of Auinterfaced with SiO2 inside water, we achieved a reliable lift-offand successful transfer/integration of vertically stacked, few-layer only 2D MoS2/WS2 vdW heterostructures on secondarysubstrates over an area of >2 × 2 cm2. Furthermore, byexploring the multifunctionalities of Au as sacrificial layers andgrowth substrates, we demonstrate large-area photoresponsive2D MoS2/SWNT vertical heterojunctions on flexible substrates.Figure 1 illustrates the sequential process for the growth and

transfer/integration of 2D MoS2/WS2 heterostructure layers asfollows: (a) On top of a cleaned SiO2/Si substrate, thin layersof Au and SiO2 are sequentially deposited followed by adeposition of W and Mo seed layers. (b) The prepared stack is

Figure 1. Illustration for the large-area growth of vertically stacked 2D MoS2/WS2 heterostructure layers on a SiO2/Au-based substrate and theirsubsequent transfer and integration in two different ways.

Figure 2. (a) Image of as-grown 2D MoS2/WS2 heterostructure layers on a SiO2/Au/SiO2/Si substrate. (b) Raman spectrum obtained from thesample, revealing the presence of both MoS2 and WS2. (c, d) Low-magnification TEM (c) and HRTEM (d) micrographs of the 2D MoS2/WS2heterostructure layers in plane view. (e) STEM-EDS elemental mapping images revealing the uniform spatial distribution of constituent elements.The scale bar is 50 nm. (f) Cross-section TEM characterizations of 2D MoS2/WS2 heterostructure layers on a SiO2 (left) and their detailedcrystalline structures (right). (g) ADF-TEM image of the corresponding MoS2/WS2 interface, revealing a distinct image contrast. (h) STEM-EDSelemental map to show the spatial localization of Mo and W at the interface.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6158

sulfurized in a low-pressure CVD chamber, which leads to achemical conversion of Mo/W to 2D MoS2/WS2 layers.Following this material growth stage, two different transfer/integration approaches can be independently pursued. In thefirst approach (Figure 1c−f). (c) A thin protective layer (PL)such as poly(methyl methacrylate) (PMMA) or polydimethyl-siloxane (PDMS) is spin-coated on top of the 2D MoS2/WS2layers. (d) The entire substrate is immersed in water, whichleads to a facile lift-off of the PL/MoS2/WS2/SiO2 layer-onlyowing to the debonding nature of Au interfaced with SiO2

(details are to be explained below). (e) The PL/MoS2/WS2/SiO2 stack is mechanically peeled off/exfoliated from thegrowth substrate. (f) The separated stack is transposed upsidedown onto a second substrate after the removal of SiO2, wherethe PL functions as an adhesion material. In the secondapproach (Figure 1g−j): (g) As-prepared 2D MoS2/WS2 layersgrown on a SiO2/Au/Si substrate are immersed in water. (h)Once removed from the water bath and air-dried, the top of thesamples is covered with a thermal release tape. (i) The tape ismechanically lifted, which leads to the separation of the 2DMoS2/WS2/SiO2 layer supported by the tape. (j) Upon the

attachment of the 2D MoS2/WS2/SiO2 layer onto a secondsubstrate, the tape is released under heating at 130 °C.38 Thisdirect growth of 2D heterostructure layers on such transferrablesubstrates offers distinct advantages over conventional 2D layertransfer methods; 2D layers can retain the structural integritiesof their as-grown states during the lift-off/transfer processessince the underlying growth substrate (SiO2 layer) itself isdirectly transferred rather than it is etched away.The quality of as-grown 2D MoS2/WS2 heterostructures was

verified by extensive structural characterizations. Figure 2ashows a photograph of as-grown 2D MoS2/WS2 on thesubstrate, exhibiting a uniform color homogeneity over theentire substrate surface of >2 × 2 cm2. Figure 2b shows theRaman spectrum collected from the same sample, exhibitingfour distinct peaks associated with the in-plane E2g

1 and theout-of-plane A1g phonon vibration modes of individual 2DMoS2 and 2D WS2 layers.

39,40 The Raman spectrum does notdisplay an indication of MoxW1−xS2 whose peaks are to bepoised in between the peaks for individual MoS2 and WS2,

41−43

which suggests that the sample is largely unalloyed within itsthickness (>10 nm). The detailed morphology of the 2D

Figure 3. (a−f) Sequential procedures for the transfer of 2D MoS2/WS2 heterostructure layers using the water assisted Au-SiO2 separation. (g)PDMS-coated 2D WS2/MoS2 heterostructure layers integrated on the surface of a cup. (h) 2D MoS2/WS2 heterostructure layers on a PDMStransferred from a thermal release tape. (i) Raman spectrum from 2D MoS2/WS2 heterostructure layers in comparison to the Raman spectra fromindividual 2D MoS2 and 2D WS2. (j) Transport characteristics of transferred 2D MoS2/WS2 heterostructure layers.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6159

MoS2/WS2 heterostructure layer was characterized by trans-mission electron microscopy (TEM). Figure 2c is a bright-fieldlow-magnification TEM micrograph of the 2D MoS2/WS2heterostructure, revealing its poly crystalline structure andcontinuous layer morphology. Figure 2d shows a plane-viewhigh-resolution TEM (HRTEM) micrograph of the samesample. The image clearly reveals Moire patterns, indicative ofthe vertical stacking of multiple 2D layers whose basal planesare misaligned with respect to the hexagonal [001] zone axis.39

Figure 2e is an energy dispersive X-ray spectroscopy (EDS)elemental mapping image of the same sample characterized in ascanning TEM (STEM) mode. The image exhibits a highlyuniform spatial distribution of all constituent components ofMo, W, and S. Furthermore, the same sample was inspected bycross-sectional TEM characterizations. Figure 2f shows TEMmicrographs of a cross-sectioned 2D MoS2/WS2 heterostruc-ture grown on top of a SiO2 layer (left). The HRTEM image(right) reveals the vertical stacking of 2D MoS2 and 2D WS2layers where each material consists of ∼10 horizontal atomiclayers with a well-defined MoS2/WS2 interface (yellow dottedline). Figure 2g shows an annular dark-field (ADF) TEMmicrograph of the corresponding MoS2/WS2 interface. Thedistinct ADF image contrast between the MoS2 and the WS2reflects the mass difference of constituting W and Mo. Theobservation is consistent with the spatial distribution of Mo andW as confirmed by the STEM-EDS elemental mapping inFigure 2h, indicating the vertical stacking of 2D MoS2 and 2DWS2 layers. All of these comprehensive structural/chemicalcharacterizations evidence the successful CVD growth of large-area, vertically stacked 2D MoS2/WS2 heterostructure layers ontop of the SiO2 deposited Au/SiO2/Si substrate. The directgrowth of 2D MoS2 (or, 2D WS2) on a secondarily SiO2 surfacehas previously been demonstrated via metal−organic CVD(MOCVD),44 which further strengthens the validity of ourapproach. Moreover, to ensure the lateral layer-by-layer growthof each constituent material avoiding the unwanted randomvertical/slanted growth of 2D layers,39 we performed our CVDgrowths with uniformly deposited metal seeds of smallthickness (typically, ∼4−6 nm). We also carried out thesequential growth of one material on the other via two-stepCVD process;28,29 that is, growth of 2D WS2 followed by 2DMoS2 or vice versa, and did not notice any significantdistinction in the resulting morphology. Details for growthconditions can be found in Method section.Figure 3a−e illustrates the transfer of CVD-grown vertically

stacked 2D MoS2/WS2 heterostructure layers, which corre-sponds to the approach in Figure 1c−f. (a) PL (PDMS in thiscase)-covered 2D MoS2/WS2 heterostructure layers grown on aSiO2/Au/SiO2/Si substrate is immersed in water. (b−c) Afterthe water immersion, the PL/MoS2/WS2/SiO2 stack becomesreadily detachable from the underlying Au layer and is ready fora lift-off after taken out of the water bath. (d) The PL/MoS2/WS2/SiO2 layer is manually peeled off from the growthsubstrate. (e) After the lift-off, the original growth substrate isfound to be still covered with the Au layer which visiblyremains intact. (f) The detached PL/MoS2/WS2 layer isobserved to be covered with SiO2 on its back side, indicatingthat the layer separation occurs at the SiO2/Au interface. Thisobservation suggests that the detached 2D MoS2/WS2heterostructure layer retains its structural integrity as it wasprotected in between the PL and the SiO2 layer during theentire lift-off process. The detached PL/MoS2/WS2 hetero-structure layer is highly flexible even with the SiO2 layer

attached on its back side (Figure 3d) and can be integratedonto secondary substrates. Figure 3g demonstrates the directintegration of the detached layer (on the surface of a cupfollowing a removal of the underlying SiO2 layer). The PL(PDMS in this case) itself functions as an adhesive materialwhich enables the sticking of the 2D WS2/MoS2 hetero-structure layer on the foreign substrates. The dotted line inFigure 3g indicates that the integrated layer well retains theoriginal shape/size of its as-lifted state which corresponds toFigure 3f. Figure 3h demonstrates the integration of another2D MoS2/WS2 heterostructure layers prepared by the transferapproach in Figure 1g−j; as-grown 2D MoS2/WS2 hetero-structure layers on Au/SiO2/Si substrate are separated by athermal release tape, following the removal of SiO2. The 2DMoS2/WS2 heterostructure layers supported by the tape areintegrated onto a foreign substrate (PDMS in this case),followed by a release of the tape under heating. The imagereveals the optical transparency of the integrated 2Dheterostructure layers similar to Figure 3g. It is worthmentioning that the complete removal of the SiO2 in theselayers (Figure 3g and h) before their integration is significantlyeasier/faster than the conventional SiO2 etching-based lift-offapproaches as the entire SiO2 surface is exposed unlike that it isembedded in between 2D heterostructure layers and Si.30 Thesuccess of this facile/reliable 2D layer separation at the SiO2/Au interface in water is attributed to the intrinsic debondingnature of Au.45−47 It has been known that the interfacialadhesion energy for Au interfaced with SiO2 is lower than mostof the noble metals. In air, the adhesion energy is ∼0.4 J/m2 forAu/SiO2, while it is >3 J/m2 for titanium (Ti), a material thathas been widely used to increase the contact adhesion of Auonto SiO2 surface. The interfacial adhesion energy is ∼1.3 J/m2

for nickel (Ni) which is also well-known for its weak adhesionto SiO2.

45,46 Moreover, it has been reported that waterpenetration drastically reduces the interfacial energies formetal/SiO2 interfaces. For instance, ∼80% lower interfacialenergy has been reported for Ni/SiO2 in water compared to inair, and this trend is valid regardless of metals interfaced withSiO2.

46 The structural integrities of 2D MoS2/WS2 hetero-structure layers detached from their growth substrates werecharacterized and confirmed by Raman spectroscopy. The blackplot in Figure 3i shows a Raman spectrum from the separated2D MoS2/WS2 heterostructure layers before the integration(Figure 3h), revealing the Raman peaks assigned to individualMoS2 and WS2. The Raman spectra (red plots) obtained fromindividual MoS2 and WS2 layers grown under the same growthcondition are presented for comparison. The Raman mappingimage (Supporting Information, Figure S1) further demon-strates the spatial uniformity of MoS2 and WS2 over a largearea. We have also characterized the Raman spectra of varioussamples before/after their lift-off (Supporting Information,Figure S2) and confirmed no significant difference in theircharacteristics. Figure 3j shows two-terminal carrier transportcharacteristics of transferred 2D MoS2/WS2 heterostructurelayers where top/bottom metal (Au) electrodes are separatelymade on each material (schematic in Figure 3j, upper inset).The plot reveals a non-Ohmic rectifying transport (also,semilogarithmic plot in Figure 3j, lower inset) confirming thepresence of 2D MoS2/WS2 heterojunctions with band offsets,consistent with previous observations.13,17,23,39,40

In addition to the functionality as a SiO2-sacrificial layer, weexplore the feasibility of Au as a growth substrate for 2D TMDlayers. To demonstrate these multifunctionalities of Au, we

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6160

directly grew 2D MoS2 layers on top of Au/SiO2/Si substratesand transfer the grown layers. Figure 4a illustrates the growth-to-transfer procedure; 2D MoS2 is grown on top of a Au/SiO2/Si substrate by the direct sulfurization of Mo, and the substrateis subsequently immersed in water. Once the substrate isremoved from the water bath and is dried, a thermal releasesupporting tape is attached to the 2D MoS2 layers on thesubstrate, and it is mechanically peeled off leading to theseparation of Au-attached 2D MoS2 from the substrate. The Auon the 2D MoS2 can be removed resulting in 2D MoS2 layersintegrated on the tape, or the entire 2D MoS2/Au layers can beintegrated onto secondary substrates. Figure 4b shows anoptical image of 2D MoS2 layers directly grown on Au-deposited SiO2/Si substrates. The 2D MoS2 layers were grownon the selected areas only where Mo films were predeposited,presenting different colors depending on the thickness of initial

Mo (3 and 10 nm for left and right, respectively). Figure 4cshows Raman characteristics from the 2D MoS2 layers directlygrown on Au-deposited substrates revealing that the intensityratio of E2g

1/A1g increases with decreasing the thickness ofinitial Mo films, consistent with previous growth studies withSiO2/Si substrates.

48 Figure 4d shows a cross-sectional ADF-TEM micrograph of a stack of 2D MoS2/Au/SiO2, and thezoomed-in image (red box) reveals that 2D MoS2 (∼8−10layers) was grown on top of Au on a large area, obtained from 3nm thick Mo deposition. Figure 4(e) is ADF-TEM micrographof a 2D MoS2/Au interface, which clearly reveals the atomicallysharp MoS2/Au interface and well-resolved 2D MoS2 layers.STEM-EDS elemental mapping analysis does not indicate anoticeable interdiffusion of Au into the 2D MoS2 layers(Supporting Information, Figure S3). Figure 4f shows a plane-view HRTEM micrograph of 2D MoS2 layers obtained from

Figure 4. (a) Illustration for the direct growth of 2D MoS2 layers on Au-deposited substrates and their subsequent transfers. (b) Images of 2D MoS2-grown on Au/SiO2/Si substrates. 2D MoS2 are selectively grown with Mo of different thicknesses (left: 3 nm, right: 10 nm). (c) Raman spectrumobtained from the 2D MoS2 layers grown on in panel b. (d) Cross-sectional ADF-TEM micrographs of 2D MoS2 grown on a Au/SiO2/Si substrate.The zoom-in image (red box) reveals the growth of continuous 2D MoS2 layers. (e) ADF-TEM micrograph revealing the sharp 2D MoS2/Auinterface. (f) Plane-view HRTEM micrograph of few-layer 2D MoS2 layers revealing Moire patterns. (g) 2D MoS2/Au before (left) and after (right)integration to a supporting tape. (h) Large-area 2D MoS2 layers attached to a supporting tape. (i) Patterned 2D MoS2 layers attached to a supportingtape.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6161

growth on Au-deposited SiO2/Si substrates (different samplefrom Figure 4d, e). The image reveals Moire patterns resultingfrom the vertical stack of multiple 2D MoS2 layers, while thefast Fourier transform (FFT) image (inset) corresponding tothe HRTEM indicates ∼3−5 layers of 2D MoS2.

49 All of thesestructural characterizations indicate that Au-deposited SiO2/Sisubstrates function as excellent growth substrates while theirpotential as sacrificial layers for facile 2D layer transfers hasalready been proved. Figure 4g shows the images of as-grown2D MoS2 on Au/SiO2/Si substrate (before) and the samesample transferred/attached to a supporting tape (after). Figure4h and i shows images of large-area (>3 cm2) and patterned 2DMoS2 attached to supporting tapes, respectively, demonstratingtheir mechanical flexibility.To verify the versatility of the Au-assisted transfer and its

applications to 2D electronics/optoelectronics, we demonstratelarge-area, heterojunction devices based on vertically stackedsingle walled carbon nanotube (SWNT) thin films and 2DMoS2/Au layers. Figure 5a illustrates a schematic for the devicefabrication; 2D MoS2/Au layers are integrated on a secondaryconductive and flexible substrate, for example, copper (Cu) foil.Subsequently, highly dispersed p-type sorted SWNTs (>99%semiconducting) are directly integrated on the surface of 2DMoS2 layers, following the selective opening of MoS2 via PDMSwindow. As a result, large-area SWNT/2D MoS2 verticalheterojunctions are realized where top/bottom electrodes aredirectly made onto the SWNT and the back side of the Cu foil,respectively (Figure 5b). Figure 5c shows a representativeimage of a large-area vertically stacked 2D MoS2/Au layersintegrated on a thin flexible Cu foil prior to the deposition ofPDMS and SWNTs. The assembled SWNT/2D MoS2 verticalheterojunctions were electrically characterized. Figure 5d showsthe current density (J)−voltage (V) characteristics of arepresentative SWNT/2D MoS2 heterojunction on a Cu foil.

Current rectification with asymmetric J−V is clearly observed(semilogarithmic plot in Figure 5d, inset), which is consistentwith the recent observation with laterally stacked SWNT/2Dmonolayer MoS2 p−n junction diodes.50 The rectification ratio,that is, the ratio of forward-to-reverse current amplitude,reaches ∼10 which is slightly smaller than the value reportedwith the micrometer-sized SWNT/2D MoS2 lateral junctionwithout gate voltages.50 The vertically stacked SWNT/2DMoS2 heterojunction devices exhibit substantive photoresponseunder an illumination of a white light. Figure 5e shows thesemilogarithmic I−V characteristics of another device (differentfrom Figure 5d) in the reverse bias regime with/without anillumination, revealing a significant photocurrent generation(i.e., > × ∼ 4−5 increase of reverse current). We investigate theorigin of the observed current rectification and photoresponseby exploring the exclusive role of the SWNT/2D MoS2heterojunction on carrier transport properties. Two-terminalcurrent−voltage (I−V) characteristics of individual SWNT filmand 2D MoS2 layers without the heterojunction in dark andunder illumination are presented (Supporting Information,Figure S4). The plots reveal symmetric I−V characteristicsfrom both the materials and some photoresponse (i.e., currentincrease of ∼1.3 times) in the 2D MoS2 with a negligiblecurrent change in the SWNT. This observation confirms thatthe current rectification and significant photocurrent observedwith the vertically stacked SWNT/2D MoS2 device indeedoriginate from the vertical SWNT/2D MoS2 heterojunction.The energy band structure of the heterojunction in Figure 5gillustrates the underlying mechanism. The band diagram isconstructed based on that the band gap energy (Eg) of p-typepure SWNT is ∼0.6 eV with the diameter of ∼1.7 nm used inthis study,51 smaller than Eg of MoS2 which is intrinsically n-type. The diagram presents a presence of type II-like band-offsets responsible for relaxing the photogenerated electron−

Figure 5. (a) Schematic for the fabrication of a vertically stacked SWNT/2D MoS2 heterojunction integrated on a flexible Cu foil. (b) Side-viewillustration of a vertically stacked SWNT/2D MoS2 heterojunction configured for electrical characterizations. (c) Image of 2D MoS2 /Au layersintegrated on a Cu foil. (d) J−V characteristics from a SWNT/2D MoS2 heterojunction showing current rectification and the correspondingsemilogarithmic presentation (inset). (e) Photoresponse characteristics from a SWNT/2D MoS2 heterojunction. (f) Energy band diagram ofSWNT/2D MoS2 heterojunction depicting the separation/diffusion of photocarriers generated under a light illumination (purple arrow). CB, VB,and Ef represent the conduction band, valence band, and Fermi energy, respectively.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6162

hole (e−−h+) pairs which subsequently diffuse across theheterojunction, resulting in photocurrent. The recombinationof charge carriers may be present caused by variousmechanisms, for example, series/interfacial resistance, limitingthe charge transport as reflected by the smaller rectificationratio compared to the previous lateral SWNT/2D MoS2heterojunction.50 It is worth emphasizing that the SWNT/2DMoS2 heterojunctions achieved in this study are significantlylarger (∼cm2) than previous studies (∼μm2) and are in avertical geometry. These attributes are more preferred for 2Doptoelectronics since the diffusion length for photogeneratedcarriers can become very small as it corresponds to the verticalheight of SNWT/2D MoS2 layers. Further optimizations ofmaterial processing and integrations are needed to realize thisadvantage.Overall, all of these demonstrations of 2D layer transfers and

their applications strengthen the generality and the versatility ofthe Au-mediated 2D layer separation benefiting from the water-assisted debonding of Au with SiO2. We also note that a longtime (>a few hours) for water immersion often results in amore spontaneous separation of 2D MoS2 layers from theunderlying Au despite their strong adhesion properties,52,53

which offers additional advantages for the deterministicseparation of 2D MoS2 layers. It is also worth mentioningthat Au has recently drawn attentions as a growth substrate forvarious 2D TMDs owing to its low reactivity with sulfur andlow solubility with metals (Mo and W).54 It has been used inthe forms of a foil37,54,55 or a deposited layer,56,57 which furtherstrengths the significance and the versatility of our study.Lastly, we demonstrate the comparison of the water-assisted

Au-mediated transfer with the conventional BOE-basedapproach to verify/highlight its intrinsic advantage of fasterand cleaner layer separation. In Figure 6, we compare thetransfer of 2D MoS2 layers grown on SiO2 (Figure 6a) and Au(Figure 6b) prepared under identical growth conditions (i.e.,identical sample size, Mo film thickness, and growth temper-ature). The results show that the Au-mediated approach takessignificantly shorter time (∼3 min vs ∼51 min) in separating2D MoS2 layers while better maintaining the structural integrityof their as-grown state.In conclusion, we have demonstrated a CVD growth of

vertically stacked 2D MoS2/WS2 vdW heterostructure layers onSiO2/Au-based substrates and their direct transfer onto foreignsubstrates facilitated by the water-assisted debonding of SiO2/Au interfaces. This uniquely combined growth/transfer strategyis highly scalable, enabling that 2D TMD layers retain theirstructural integrity up to a centimeter-scale. As a proof-of-concept, large-scale SWNT/2D MoS2 vertical heterojunctionsare presented, which are difficult to achieve with conventional2D layer stacking approaches. This study is believed to

accelerate the exploration of 2D TMD heterostructure layersfor device building-blocks in emerging flexible electronics andoptoelectronics.

Methods. CVD Growth. A commercial SiO2/Si wafer is cutinto ∼2 × 2 cm2 pieces followed by sequential cleaning withacetone and methanol. The wafer pieces are sequentiallydeposited with Au (∼50−100 nm), SiO2 (∼300−450 nm), andmetals (Mo and W: ∼4−6 nm) by electron beam evaporation(Temescal FC-2000 evaporator). The prepared substrates areloaded into a quartz tube CVD furnace which is pumped downto <1 mTorr. After purging with argon (Ar) gas, the furnace isheated up to ∼650−700 °C with a flow of Ar [100 standardcubic centimeters per minute (SCCM)] at an operatingpressure of ∼100 mTorr. After ∼30 min reaction, the furnaceis naturally cooled down, and the substrate is taken out of thefurnace. It is observed that the color of the substrate changesfrom silver to dark green, which indicates the sulfurization ofMo and W. Growth conditions for 2D MoS2/WS2 on SiO2/Ausubstrates are identical for those for 2D MoS2 on Au.

Lift-off and Transfer. In Figure 3, 2D MoS2/WS2 layersgrown on SiO2/Au-based substrates with (Figure 3g) orwithout (Figure 3h) PL are immersed in water. Prior to theimmersion, small areas on the corners of the samples areexposed by mechanical scratch, which is to facilitate the waterpenetration into the Au/SiO2 interface. After the waterimmersion, the samples are taken out of the water bath, andthe samples are mechanically separated by using a tweezer.Before the 2D layer integrations in Figure 3g and h, theexposed SiO2 on the samples is removed by BOE. In Figure 4, athermal release tape (Semiconductor Equipment Corp.) isdirectly attached to as-grown 2D MoS2 layers on Au-depositedsubstrates following water immersion. The tape with the 2DMoS2/Au layers is manually detached from the growthsubstrate upon heating at 130 °C at a hot plate. Au residualson the back side of the MoS2 can be removed by Au etchant(iron chloride, FeCl3).

Raman and TEM Characterizations. Raman spectra werecollected using Almega XR Raman spectrometer equipped withan Olympus BX51 microscope at a laser wavelength of 532 nm.The crystalline structure and the chemical composition of as-grown 2D layers were characterized using a JEOL ARM200FFEG-TEM/STEM with a Cs-corrector. STEM-EDS analysiswas carried out using EDAX detector (SDD type 80T) and ananalysis software (AZtecTEM, Oxford). All TEM/STEMoperations were performed at an accelerating voltage of 200kV. Cross-sectional TEM samples were prepared by focusedion beam (FIB; Quanta 2D FEG, FEI) based milling and lift-out techniques. As-grown 2D TMDs were deposited with ∼100nm thick carbon (C) and platinum (Pt) layers and weresubsequently cross-sectioned inside a FIB via gallium (Ga) ion

Figure 6. (a) BOE-based separation of 2D MoS2−PMMA layer which took ∼51 min (left), resulting in partially peeled-off 2D MoS2 layers (right).(b) Water-based separation of 2D MoS2/Au-PMMA layer which took ∼3 min (left), resulting in clean transfer (right). Both of the samples wereprepared under identical experimental conditions.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6163

milling at 30 keV. The prepared TEM specimen was placedonto a Cu TEM grid with a micromanipulator (Omniprobe)inside the FIB.Electrical and Photoresponse Characterizations. 2D

MoS2/Au layers transferred to a Cu foil were selectivelycovered with a PDMS window. Diluted p-type SWNTs insolution were deposited on the exposed 2D MoS2 extending tothe surface of PDMS window. Silver (Ag) paste is directlyapplied to the SWNT on PDMS for a top contact, and thebottom contact is connected to the back side of Cu foil.Electrical and photoresponse characterizations were performedat room temperature in Micromanipulator 6200 probe stationusing a semiconductor parameter analyzer (KEYSIGHTB1500A). The device is globally illuminated with a whitelight source (intensity: 20 W/m2), and output characteristicswere obtained before and after illumination.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge via theInternet at The Supporting Information is available free ofcharge on the ACS Publications website at DOI: 10.1021/acs.nanolett.7b02776.

Additional TEM data, Raman data, and I−V measure-ment data (PDF)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: hschung13@kbsi.re.kr.*E-mail: yeonwoong.jung@ucf.edu.

ORCIDTania Roy: 0000-0003-1131-8068Yeonwoong Jung: 0000-0001-6042-5551Author ContributionsY.J. conceived the idea, directed the project, and organized themanuscript. H.-S.C. supervised and performed all of the TEMexperiments and analyzed the TEM data with the help of J.H.K.and K.H.O. M.A.I. synthesized and transferred the 2Dmaterials, conducted Raman characterizations, and fabricateddevices with the help of A.S., N.C., D.W., and S.I.K. M.A.I.performed the electrical and photoresponse characterizationswith the help of H.K. and T.R. All of the authors contributed tothe discussion of the paper and approved the manuscript.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSY.J. acknowledges a start-up fund from the University ofCentral Florida and the use of Materials CharacterizationFacility, AMPAC, at the University of Central Florida. H.-S.C.was supported by National Research Foundation of Korea(NRF) grant funded by Korea government (MSIP) (no.2015R1C1A1A01052727). K.H.O. acknowledges the financialsupport of National Research Foundation (NRF-2016M3C1B5906481) and Ministry of Public Safety andSecurity (MPSS-CG-2016-02) from Korean government.

■ REFERENCES(1) Novoselov, K. S.; Mishchenko, A.; Carvalho, A.; Castro Neto, A.H. Science 2016, 353, aac9439.

(2) Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J.; Strano, M.S. Nat. Nanotechnol. 2012, 7, 699−712.(3) Bhimanapati, G. R.; Lin, Z.; Meunier, V.; Jung, Y.; Cha, J.; Das,S.; Xiao, D.; Son, Y.; Strano, M. S.; Cooper, V. R.; et al. ACS Nano2015, 9, 11509−11539.(4) Akinwande, D.; Petrone, N.; Hone, J. Nat. Commun. 2014, 5,5678.(5) Geim, A. K.; Grigorieva, I. V. Nature 2013, 499, 419−425.(6) Lotsch, B. V. Annu. Rev. Mater. Res. 2015, 45, 85−109.(7) Li, M.; Chen, C.; Shi, Y.; Li, L. Mater. Today 2016, 19, 322−335.(8) Ajayan, P. M.; Kim, P.; Banerjee, K. Phys. Today 2016, 69, 38−44.(9) Pant, A.; Mutlu, Z.; Wickramaratne, D.; Cai, H.; Lake, R. K.;Ozkan, C.; Tongay, S. Nanoscale 2016, 8, 3870−3887.(10) Chen, K.; Wan, X.; Xu, J. Adv. Funct. Mater. 2017, 27, 1603884.(11) Castellanos-Gomez, A.; Buscema, M.; Molenaar, R.; Singh, V.;Janssen, L.; van der Zant, H. S. J.; Steele, G. A. 2D Mater. 2014, 1,011002.(12) Nourbakhsh, A.; Zubair, A.; Dresselhaus, M. S.; Palacios, T.Nano Lett. 2016, 16, 1359−1366.(13) Lee, C.; Lee, G.; van der Zande, A. M.; Chen, W.; Li, Y.; Han,M.; Cui, X.; Arefe, G.; Nuckolls, C.; Heinz, T.; et al. Nat. Nanotechnol.2014, 9, 676−681.(14) Fang, H.; Battaglia, C.; Carraro, C.; Nemsak, S.; Ozdol, B.;Kang, J. S.; Bechtel, H. A.; Desai, S. B.; Kronast, F.; Unal, A. A.; et al.Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 6198−6202.(15) Furchi, M. M.; Pospischil, A.; Libisch, F.; Burgdorfer, J.; Mueller,T. Nano Lett. 2014, 14, 4785−4791.(16) Chiu, M.; Zhang, C.; Shiu, H.; Chuu, C.; Chen, C.; Chang, C.S.; Chen, C.; Chou, M.; Shih, C.; Li, L. Nat. Commun. 2015, 6, 7666.(17) Hong, X.; Kim, J.; Shi, S. F.; Zhang, Y.; Jin, C.; Sun, Y.; Tongay,S.; Wu, J.; Zhang, Y.; Wang, F. Nat. Nanotechnol. 2014, 9, 682.(18) Tongay, S.; Fan, W.; Kang, J.; Park, J.; Koldemir, U.; Suh, J.;Narang, D. S.; Liu, K.; Ji, J.; Li, J.; Sinclair, R.; Wu, J. Nano Lett. 2014,14, 3185−3190.(19) Ceballos, F.; Bellus, M. Z.; Chiu, H.; Zhao, H. ACS Nano 2014,8, 12717−12724.(20) Rivera, P.; Schaibley, J. R.; Jones, A. M.; Ross, J. S.; Wu, S.;Aivazian, G.; Klement, P.; Seyler, K.; Clark, G.; Ghimire, N. J.; et al.Nat. Commun. 2015, 6, 6242.(21) Cha, S.; Cha, M.; Lee, S.; Kang, J.; Kim, C. Sci. Rep. 2016, 5,17877.(22) Ma, D.; Shi, J.; Ji, Q.; Chen, K.; Yin, J.; Lin, Y.; Zhang, Y.; Liu,M.; Feng, Q.; Song, X.; Guo, X.; Zhang, J.; Zhang, Y.; Liu, Z. Nano Res.2015, 8, 3662.(23) Gong, Y.; Lin, J.; Wang, X.; Shi, G.; Lei, S.; Lin, Z.; Zou, X.; Ye,G.; Vajtai, R.; Yakobson, B. I.; et al. Nat. Mater. 2014, 13, 1135−1142.(24) Yu, Y.; Hu, S.; Su, L.; Huang, L.; Liu, Y.; Jin, Z.; Purezky, A. A.;Geohegan, D. B.; Kim, K. W.; Zhang, Y.; Cao, L. Nano Lett. 2015, 15,486−491.(25) Heo, H.; Sung, J. H.; Jin, G.; Ahn, J.; Kim, K.; Lee, M.; Cha, S.;Choi, H.; Jo, M. Adv. Mater. 2015, 27, 3803−3810.(26) Lin, Y.; Ghosh, R. K.; Addou, R.; Lu, N.; Eichfeld, S. M.; Zhu,H.; Li, M.; Peng, X.; Kim, M. J.; Li, L.; et al. Nat. Commun. 2015, 6,7311.(27) Gong, Y.; Lei, S.; Ye, G.; Li, B.; He, Y.; Keyshar, K.; Zhang, X.;Wang, Q.; Lou, J.; Liu, Z.; et al. Nano Lett. 2015, 15, 6135−6141.(28) Xue, Y.; Zhang, Y.; Liu, Y.; Liu, H.; Song, J.; Sophia, J.; Liu, J.;Xu, Z.; Xu, Q.; Wang, Z.; et al. ACS Nano 2016, 10, 573−580.(29) Wu, C.; Chang, X.; Chu, T.; Chen, H.; Wu, C.; Lin, S. NanoLett. 2016, 16, 7093−7097.(30) Yu, J. H.; Lee, H. R.; Hong, S. S.; Kong, D.; Lee, H.; Wang, H.;Xiong, F.; Wang, S.; Cui, Y. Nano Lett. 2015, 15, 1031−1035.(31) Jung, Y.; Shen, J.; Sun, Y.; Cha, J. J. ACS Nano 2014, 8, 9550−9557.(32) Phan, H. D.; Kim, Y.; Lee, J.; Liu, R.; Choi, Y.; Cho, J. H.; Lee,C. Adv. Mater. 2017, 29, 1603928.(33) Li, H.; Wu, J.; Huang, X.; Yin, Z.; Liu, J.; Zhang, H. ACS Nano2014, 8, 6563−6570.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6164

(34) Gurarslan, A.; Yu, Y.; Su, L.; Yu, Y.; Suarez, F.; Yao, S.; Zhu, Y.;Ozturk, M.; Zhang, Y.; Cao, L. ACS Nano 2014, 8, 11522−11528.(35) Van Ngoc, H.; Qian, Y.; Han, S. K.; Kang, D. J. Sci. Rep. 2016, 6,33096.(36) Lu, Z.; Sun, L.; Xu, G.; Zheng, J.; Zhang, Q.; Wang, J.; Jiao, L.ACS Nano 2016, 10, 5237−5242.(37) Yun, S. J.; Chae, S. H.; Kim, H.; Park, J. C.; Park, J.; Han, G. H.;Lee, J. S.; Kim, S. M.; Oh, H. M.; Seok, J.; et al. ACS Nano 2015, 9,5510−5519.(38) Lin, Z.; Zhao, Y.; Zhou, C.; Zhong, R.; Wang, X.; Tsang, Y. H.;Chai, Y. Sci. Rep. 2016, 5, 18596.(39) Choudhary, N.; Park, J.; Hwang, J. Y.; Chung, H.; Dumas, K. H.;Khondaker, S. I.; Choi, W.; Jung, Y. Sci. Rep. 2016, 6, 25456.(40) Woods, J. M.; Jung, Y.; Xie, Y.; Liu, W.; Liu, Y.; Wang, H.; Cha,J. J. ACS Nano 2016, 10, 2004−2009.(41) Lui, C. H.; Malard, L. M.; Kim, S.; Lantz, G.; Laverge, F. E.;Saito, R.; Heinz, T. F. Nano Lett. 2012, 12, 5539−5544.(42) Zhang, M.; Wu, J.; Zhu, Y.; Dumcenco, D. O.; Hong, J.; Mao,N.; Deng, S.; Chen, Y.; Yang, Y.; Jin, C.; et al. ACS Nano 2014, 8,7130−7137.(43) Chen, Y.; Xi, J.; Dumcenco, D. O.; Liu, Z.; Suenaga, K.; Wang,D.; Shuai, Z.; Huang, Y.; Xie, L. ACS Nano 2013, 7, 4610−4616.(44) Kang, K.; Xie, S.; Huang, L.; Han, Y.; Huang, P. Y.; Mak, K. F.;Kim, C.; Muller, D.; Park, J. Nature 2015, 520, 656−660.(45) Kennedy, M. S.; Moody, N. R.; Adams, D. P.; Clift, M.; Bahr, D.F. Mater. Sci. Eng., A 2008, 493, 299−304.(46) Lee, C. H.; Kim, J.; Zou, C.; Cho, I. S.; Weisse, J.; Nemeth, W.;Wang, Q.; van Duin, A.; Kim, T.; Zheng, X. Sci. Rep. 2013, 3, 2917.(47) Zhao, J.; Guo, Y.; Cai, L.; Li, H.; Wang, K. X.; Cho, I. S.; Lee, C.H.; Fan, S.; Zheng, X. ACS Energy Lett. 2016, 1, 68−75.(48) Jung, Y.; Shen, J.; Liu, Y.; Woods, J. M.; Sun, Y.; Cha, J. J. NanoLett. 2014, 14, 6842−6849.(49) Jeon, J.; Jang, S. K.; Jeon, S. M.; Yoo, G.; Jang, Y. H.; Park, J. H.;Lee, S. Nanoscale 2015, 7, 1688−1695.(50) Jariwala, D.; Sangwan, V. K.; Wu, C. C.; Prabhumirashi, P. L.;Geier, M. L.; Marks, T. J.; Lauhon, L. J.; Hersam, M. C. Proc. Natl.Acad. Sci. U. S. A. 2013, 110, 18076−18080.(51) Shekhar, S.; Stokes, P.; Khondaker, S. I. ACS Nano 2011, 5 (3),1739−1746.(52) Zhao, J.; Yu, H.; Chen, W.; Yang, R.; Zhu, J.; Liao, M.; Shi, D.;Zhang, G. ACS Appl. Mater. Interfaces 2016, 8, 16546−16550.(53) Desai, S. B.; Madhvapathy, S. R.; Amani, M.; Kiriya, D.; Hettick,M.; Tosun, M.; Zhou, Y.; Dubey, M.; Ager, J.; Chrzan, D.; et al. Adv.Mater. 2016, 28, 4053−4058.(54) Gao, Y.; Liu, Z.; Sun, D.; Huang, L.; Ma, L.; Yin, L.; Ma, T.;Zhang, Z.; Ma, X.; Peng, L.; Cheng, H.; Ren, W. Nat. Commun. 2015,6, 8569.(55) Shi, J.; Ma, D.; Han, G.; Zhang, Y.; Ji, Q.; Gao, T.; Sun, J.; Song,X.; Li, C.; Zhang, Y.; Lang, X.; Zhang, Y.; Liu, Z. ACS Nano 2014, 8,10196−10204.(56) Bradley, A. J.; M. Ugeda, M. M.; da Jornada, F.; Qiu, D. Y.;Ruan, W.; Zhang, Y.; Wickenburg, S.; Riss, A.; Lu, J.; Mo, S.; et al.Nano Lett. 2015, 15, 2594−2599.(57) Grønborg, S. S.; Ulstrup, S.; Bianchi, M.; Dendzik, M.; Sanders,C. E.; Lauritsen, J. V.; Hofmann, P.; Miwa, J. A. Langmuir 2015, 31,9700−9706.

Nano Letters Letter

DOI: 10.1021/acs.nanolett.7b02776Nano Lett. 2017, 17, 6157−6165

6165