Tuning Optical Signatures of Single- and Few-Layer MoS2 by Blown

Jun 19, 2017 - As the strain increases via bulging, we achieve continuous tuning of Raman and photoluminescence (PL) signatures in single- and few-lay...
1 downloads 6 Views 2MB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Communication

Tuning Optical Signatures of Single- and Few-Layer MoS by Blown-Bubble Bulge Straining up to Fracture 2

Rui Yang, Jaesung Lee, Souvik Ghosh, Hao Tang, R. Mohan Sankaran, Christian A. Zorman, and Philip X.-L. Feng Nano Lett., Just Accepted Manuscript • Publication Date (Web): 19 Jun 2017 Downloaded from http://pubs.acs.org on June 19, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Tuning Optical Signatures of Single- and Few-Layer MoS2 by Blown-Bubble Bulge Straining up to Fracture Rui Yang1, Jaesung Lee1, Souvik Ghosh2, Hao Tang1, R. Mohan Sankaran2, Christian A. Zorman1, Philip X.-L. Feng1,* 1

Electrical Engineering, 2Chemical and Biomolecular Engineering, Case School of Engineering, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106, USA

Emerging atomic layer semiconducting crystals such as molybdenum disulfide (MoS2) are promising candidates for flexible electronics and strain-tunable devices due to their ultrahigh strain limits (up to ~20-30%) and strain-tunable bandgaps. However, high strain levels, controllable isotropic and anisotropic biaxial strains in single- and few-layer MoS2 on device-oriented flexible substrates permitting convenient and fast strain tuning, remain unexplored.

Here, we demonstrate a ‘blown-bubble’ bulge technique for efficiently

applying large strains to atomic layer MoS2 devices on flexible substrate.

As strain

increases via bulging, we achieve continuous tuning of Raman and photoluminescence (PL) signatures in single- and few-layer MoS2, including splitting of Raman peaks. With proper clamping of the MoS2 crystals, we apply up to ~9.4% strain in the flexible substrate, which −2.6% strain measured by PL, causes a doubly clamped single-layer MoS2 to fracture at 2.2− and 2.9− −3.5% strain measured by Raman spectroscopy. This study opens new pathways for exploiting 2D semiconductors on stretchable substrates for flexible electronics, mechanical transducers, tunable optoelectronics, and biomedical transducers on curved and bulging surfaces. KEYWORDS: 2D Semiconductors, Atomic Layer MoS2, Bulge Test, Strain Tuning, Raman Spectroscopy, Photoluminescence (PL) *

Corresponding Author. Email: [email protected]. -1-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Two-dimensional (2D) semiconductors such as atomic layers of transition metal dichalcogenides (TMDCs) have excited great interest toward flexible and transparent electronics because of their high mobilities maintained at atomic-scale thickness, exceptional flexibility and stretchability, and sizable bandgaps.1,2,3,4 As a forerunner and representative TMDC, molybdenum disulfide (MoS2) is a layered semiconductor with strain-tunable band structure and electron mobility,5,6,7,8,9,10 making it highly suitable for foldable and stretchable electronics, mechanical transducers and strain-tunable optoelectronics.11,12,13,14 In order to use MoS2 in flexible devices, quantitative understanding of its material properties under varying strain till its fractures is necessary. Different methods have already been attempted to strain 2D semiconductors, and among them, uniaxial straining techniques are mainly based on bending of the polymeric substrate as a beam,5,6,7,12,13,14,15, 16,17,18 and biaxial straining is either based on piezoelectric effect in the substrate19 or thermal expansion of the substrate.20 The shifts of photoluminescence (PL) peaks,5,6 Raman modes,21,22 and changes in transistor performance,13 as well as nanomechanical resonances 23 with strain in MoS2 structures have been characterized. There have also been measurements of Raman and PL peak shifts that revealed biaxial tensile strain caused by the mismatch in thermal expansion coefficients between MoS2 and SiO2/Si substrate during chemical vapor deposition (CVD) growth. This biaxial strain, however, was predefined by the substrate and growth, thus may not be readily tuned or controlled once the CVD growth is finished.24 In situ, continuous monitoring of Raman and PL characteristics of atomic layer MoS2 crystals under controllable biaxial strain levels on flexible substrates and in device-oriented platforms have not yet been demonstrated, and detailed device characteristics of MoS2 at large strain until fracture remain to be explored.

-2-

ACS Paragon Plus Environment

Page 2 of 24

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Bulge measurement is an effective method established for characterizing mechanical properties of thin film materials until their fracture, with a wide range of applications in research and industry. A typical configuration includes creating a differential pressure on two sides of the thin film to bulge it, while monitoring the film’s load-deflection behavior, and then extracting the stress-strain relationship. Starting 1959, bulge measurement has been performed on gold and silver films25 and then on polyimide films,26 where the Young’s modulus, residual stress and fracturing strain are measured. Later this method has been further developed and extended to other technologically important materials such as thin copper (Cu) films that are key to metallic interconnects in integrated circuits (IC), 27,28 as well as silicon nitride (SiNx), 29,30 polysilicon (poly-Si),31 and silicon carbide (SiC) thin films32 that are the major structural materials in both IC and microelectromechanical systems (MEMS). Bulging of polyimide and Cu films have led to fracturing of the materials, both at approximately 4% strain,26,27 while bulge and tensile measurements of poly-Si, SiNx and SiC have demonstrated fracturing strain levels at 0.7−1.7%, 0.3−2.3%, and 0.1−0.4%, respectively.30,33,34,35,36,37 Although recently it has been reported that quite high strain can be achieved in fully clamped suspended MoS2 by pressurizing the cavity sealed by the MoS2 membrane,38 this method appears to rely on very slow gas diffusion through the SiO2 layer to pressurize the cavity, which usually requires days to build up the necessary pressure difference for straining. Therefore, facile, fast and efficient straining techniques that are device-oriented and compatible with flexible substrate are highly desirable. Today, it is highly intriguing to explore the bulging of 2D semiconductors on various substrates to attain high strain levels simply and quickly in device platforms compatible with mainstream MEMS bulging and industrial protocols.

Moreover, when combined with flexible substrates such as

polydimethylsiloxane (PDMS), we envision and demonstrate here that the conventional bulge

-3-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

testing becomes a ‘blown-bubble’ bulge technique, featuring much boosted deformations. In addition to this blown-bubble bulge from planar to hemisphere, we note that another relevant method of blowing flexible polymer ‘balloons’ has been exploited to strain, stretch, and align nanostructures adhered on the balloon surface, so as to fabricate large arrays of aligned silicon nanowires and nanotubes. 39 , 40 , 41 , 42 , 43 By employing the blown-bubble method to bulge the polymer substrate to efficiently strain the MoS2 crystals on top, here we attain continuous and broad tuning of the optical signatures of 2D TMDCs on flexible substrates. The amount of strain can be continuously tuned by changing the gas pressure underneath the PDMS, and the gas pressure and thus the strain can be changed conveniently and quickly (in seconds). In this study, we first strain single- and few-layer MoS2 structures until they fracture using the novel blown-bubble bulge technique, on flexible PDMS substrate. The PDMS is mounted onto a specially-designed holder, and the differential gas pressure or amount of strain on PDMS can be applied fast (in seconds) with precise control. The MoS2 flakes are clamped with silicone adhesive to prevent slippage during straining, while unclamped devices are also measured in order to evaluate the effectiveness of this anchoring approach. We continuously record the optical images and the signatory Raman and PL spectra with increasing strain until the MoS2 structures slide and finally fracture, for single- and few-layer MoS2 devices. For single-layer MoS2, PL peak shifts linearly by -41±3 meV/% strain, and Raman peak E12g shifts by -2.1±0.2 cm-1/% strain, and the MoS2 structure fractures at 2.2−2.6% strain measured with PL, and 2.9−3.5% strain measured with Raman. Though the MoS2 is only clamped at two sides, the strain is isotropic biaxial strain at low strain levels below 1.2%, confirmed by the linear shift of optical signatures. Above this level, the strain becomes anisotropic due to sliding at larger strain (1.2% to 7% strain in PDMS). At large strain, splitting of the Raman peak E12g is observed, -4-

ACS Paragon Plus Environment

Page 4 of 24

Page 5 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

because strain breaks the symmetry of the lattice. At 7% to 9.4% strain in PDMS, the MoS2 structure fractures.

While other methods of straining MoS2 have been attempted, we

demonstrate here the blown-bubble bulge technique with flexible substrates, which is not only convenient and fast to achieve large strain until MoS2 fractures, but also is carried out in deviceoriented configurations, which is realistic and pertinent to many device applications such as flexible electronics, optoelectronics, and biomedical transducers on curved substrates.

The

blown-bubble bulge technique is also readily applicable to other 2D semiconductor devices for studying their strain responses. Blown-Bubble Bulge Characterization Apparatus.

Figure 1a illustrates the experimental

configuration for applying strain to MoS2. We change the pressure applied at the bottom surface of the PDMS that the MoS2 structure is adhered to, while the top surface is kept at atmosphere pressure, thereby bulging up the PDMS diaphragm. The sample is mounted on a custom-built chuck that enables N2 gas pressure to be applied to the back side of PDMS via a gas manifold and a pressure controller, thus causing it to bulge as shown in Figure 1b and 1c, and the pressure stabilizes in seconds. The PDMS sheet is clamped firmly to the chuck by a specially-designed circular acrylic holder with a circular opening in the middle that serves to define the bulging region. Prior to mounting, MoS2 flakes are mechanically exfoliated onto PDMS, then thin flakes are identified under an optical microscope and their thicknesses are confirmed by their Raman44 and PL45,46 signatures. To prevent sliding of the MoS2 flake during the bulging and straining measurements, the MoS2 is clamped on the PDMS with silicone adhesive, as shown in Figure 1b. Upon the application of pressure, the PDMS bulges upward, and the MoS2 attached to the PDMS surface is subjected to a tensile strain, which causes changes in the MoS2 band structure and phonon modes, represented by shifts in PL and Raman peaks. The excitation laser wavelength

-5-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 24

for Raman and PL measurement is 532 nm, with the laser power incident on the MoS2 structure typically within l00−200 µW to minimize laser heating effects while providing enough signal to noise ratio to resolve the peak shifts.

The Raman and PL signals are collected with a

Spectrometer and then a CCD detector. The setup is versatile and can be extended to study the strain tuning of other 2D semiconductors with different geometry. At high differential pressures, the PDMS bulges as a blown bubble, as shown in the optical image in Figure 1c. Finite element modeling (FEM) in Figure 1d as well as analytical solution is used to discern the PDMS deflection and strain at a given pressure. A photograph of a representative clamped bi-layer (2L) MoS2 structure on PDMS is shown in Figure 1e. Acrylic Holder

MoS2

Bulging PDMS

Clamp

PDMS

(b)

Gas Inlet

(a)

(c)

24kPa (3.5psi)

5mm

Silicone Clamp PDMS 2L MoS2 on Top of PDMS

(d)

Figure 1.

(e)

10µm

Experimental setup and representative MoS2 device for bulging measurement.

(a) 3D

illustration of the experimental setup with gas inlet and outlet (covered by the PDMS) on the chuck, and the acrylic holder for fixing the PDMS. (b) Illustration of the cross section of the MoS2 device, showing -6-

ACS Paragon Plus Environment

Page 7 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

that the MoS2 is clamped with silicone adhesive, and strained on top of the bulging PDMS.

(c)

Photograph of the bulging PDMS observed during experiment. (d) FEM simulation of the PDMS deformation under a differential gas pressure of 2.4×104 Pa (3.5 psi). (e) Optical image of a clamped 2L MoS2 device, with the yellow dashed lines showing the boundary of the clamp and the white dashed lines showing the boundary of the 2L MoS2.

Device Responses to Bulge at Varying Pressures. Representative optical images of a MoS2 structure on PDMS are shown in Figure 2a (before clamping) and Figure 2b (after clamping), which contains both single-layer (1L) and tri-layer (3L) regions. As the differential pressure is gradually ramped up we simultaneously record the PL and Raman spectra. In the 1L region, as we increase the pressure, we observe a continuous redshift of the direct transition PL peak A and Raman peak E12g as shown in Figure 2c and 2d, confirming that strain is applied to MoS2 by bulging the PDMS. The tendency of redshift and the decrease of PL intensity with tensile strain in MoS2 is in accordance with previous reports.5,6 The largest PL peak shift observed is ~90 meV, and the greatest Raman peak shift is ~6 cm-1 for the E12g mode. We observe in Figure 2d that the Raman peak also shows redshift with strain, and the E12g peak clearly splits into two peaks when the strain is large because the strain modifies the lattice structure and breaks the symmetry, which is in accordance with previous report.6 The Raman peak splitting indicates that the strain is anisotropic at large strain, due to the sliding between MoS2 and the flexible substrate, especially in the unclamped direction. The new observations of both the isotropic biaxial strain and the anisotropic strain at different bulging stages, plus the transition phenomena, are interesting and have implications to structures and devices with asymmetric geometries. We note the two split peaks as E1+2g and E1-2g, and fit the curves to a Lorentzian function to extract -7-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

their peak positions. The A1g Raman mode is the out-of-plane phonon mode, and has relatively small peak position shift with strain compared with E12g mode, which agrees with previous reports.21 In the 3L region, we observe redshift of both PL peak A (direct transition) and I (indirect transition), as well as Raman peak E12g, and the amount of the shift is smaller than the 1L region, as shown in Figure 2e and 2f. This is in accordance with previous reports that the responsivity of peak position shift for 3L is smaller than 1L region.47

-8-

ACS Paragon Plus Environment

Page 8 of 24

Page 9 of 24

As-Exfoliated MoS2 3L 3L

1L

1L Silicone Clamp 10µm

(a)

6.89kPa (1psi)

Intensity (a.u.)

Intensity (a.u.)

1L Region

10.34kPa (1.5psi)

4.83kPa (0.7psi) 3.45kPa (0.5psi) 2.07kPa (0.3psi)

10000 0.69kPa (0.1psi)

500

1.7

1.8

1.9

0.69kPa

3000

3.45kPa 2.07kPa

5x

370

380

2000

E12g

A1g

6.89kPa 4.83kPa 3.45kPa

0.69kPa

no strain 360

1.8

420

10.34kPa

no strain 1.6

410

13.79kPa

2.76kPa 2.07kPa 0.69kPa

370

(f)

Energy (eV)

400

17.24kPa

1000

1.38kPa

390

Raman Shift (cm-1)

24.13kPa

Intensity (a.u.)

4000

1.4

2.07kPa

A

24.13kPa 17.24kPa 10.34kPa

4.83kPa

0

3.45kPa 2.76kPa

360

5.52kPa

2000

4.83kPa

(d)

Energy (eV)

6000

Intensity (a.u.)

1500

6.89kPa

no strain

I

(e)

2000

no strain

(c)

A1g

2500 13.79kPa 10.34kPa

1000

1.38kPa (0.2psi)

0 1.6

E12g

3000 24.13kPa 17.24kPa

17.24kPa (2.5psi)

20000

10µm

(b)

A

24.13kPa (3.5psi) 30000

3L Region

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

380

390

400

410

420

-1

Raman Shift (cm )

Figure 2. Strain tuning of 1L and 3L MoS2. (a) & (b) Optical images showing the MoS2 flake (a) after exfoliation onto PDMS, and (b) after being clamped with silicone adhesive, with the boundary of the clamp and the MoS2 flake outlined by the yellow and white dashed lines, respectively. (c) & (d) -9-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

Characterization of the 1L region as indicated in (a) under different bulging and straining conditions, using (c) PL and (d) Raman spectroscopy. The differential pressure on PDMS is indicated on top of the corresponding spectrum. (e) & (f) Characterization of the 3L region as indicated in (a) under different bulge levels, using (e) PL and (f) Raman spectroscopy. The indirect transition (I) peak data in (e) are magnified for clarity. The spectra are vertically shifted, and the red dashed lines are the guidance to show the trend of peak shift.

In order to carefully examine how the strain in MoS2 correlates with the applied bulging of the PDMS, we first calibrate the displacement near the center of the PDMS (h) under different pressures, as shown in Figure 3a. This is achieved by measuring the change in focus at different pressures using an optical microscope. The displacement is also simulated with FEM, which shows a similar trend as the measured displacement, while the discrepancy could be attributed to the non-ideal clamping conditions. The displacement is further converted to strain in the PDMS using Hencky’s model for large deflection of circular membranes, 48 using the radius of the acrylic clamp opening a=3.2 mm, thickness of PDMS t=432 µm, PDMS Young’s modulus EY=1 MPa, PDMS Poisson’s ratio ν=0.5, and assuming there is no initial tension in the PDMS film. Note that here we calculate the strain in the PDMS without considering the local effect of the MoS2. Since MoS2 has a much higher Young’s modulus (270 GPa) than PDMS, although MoS2 thickness (0.7 nm for single-layer) is much smaller than the PDMS, it could still cause a deviation in the local strain as compared to the rest of the PDMS substrate. Nevertheless, this calculation provides a reasonable estimate of the strain in PDMS (εPDMS) near the MoS2 structure, and the strain in MoS2 (εMoS2) will be determined separately, as we will discuss later.

-10-

ACS Paragon Plus Environment

Page 11 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Details about the Hencky’s model and calculation of strain in the PDMS membrane are described in Supporting Information. We obtain nearly linear relationship between the applied pressure and the strain in PDMS, which is ~0.3% strain in PDMS per kilopascal of differential pressure applied. To determine the strain actually applied to the MoS2, we summarize in Figure 3 the Raman and PL peak position shifts with strain in the PDMS for the 1L region of the same MoS2 structure as shown in Figure 2. Figure 3c shows that in the range of 0% to 1.2% strain (green area), the PL peak shifts linearly with responsivity of -41±3 meV/% strain, and Figure 3d shows that the Raman peak E1-2g shifts with responsivity of -2.1±0.2 cm-1/% strain. We use the E1-2g peak instead of E1+2g peak to calculate the strain responsivity and strain in MoS2 because it shows more clear trend and larger shift, and calculation using E1+2g peak results in different strain responsivity but similar strain level in MoS2. The A1g Raman mode shifts by -0.5±0.1 cm1

/% strain in the same range of strain. In this range, the strain in MoS2 is considered biaxial

because the peak shifts linearly with strain, indicating no sliding. This is also confirmed by reversible bulging measurement of another device, and the peak shifts back when the strain is released, which is repeatable (see Supporting Information Figure S2). Above 1.2% and below 7% strain (as shown by the pink area in Figure 3c and 3d), we observe a change in the slope of the PL and Raman peak shift with strain, which indicates sliding between the MoS2 structure and the PDMS substrate. Since MoS2 should slide more in the unclamped direction, the strain is anisotropic, thus by changing the strain level, we achieve both isotropic biaxial and anisotropic strain. Though there is sliding, the net strain is still sufficient to shift the PL and Raman peaks. If we assume the MoS2 PL and Raman peak responsivity with strain is still the same as the small strain region, then we can calculate the total strain in the MoS2 from the peak position shifts, -11-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

which results in ~1% additional strain in MoS2 from PL data, and ~1.6% additional strain in MoS2 from Raman measurement, corresponding to a 5.8% increase of strain in the PDMS. At ~7-9.4% strain in the PDMS, the MoS2 structure begins to fracture as shown by the red dashed line and the red arrow in Figure 3b, and the PL and Raman peak positions shift backwards (blueshifts), but not to the original values before straining, suggesting that without clamping, there is still a certain amount of tensile strain induced in the MoS2, or the MoS2 forms some wrinkles due to previous bulging, which induces local strain. We find that the cracks align mainly with the clamping direction, which suggests that close to the fracture strain, the strain in the MoS2 is mostly in the clamped direction, while in other directions the MoS2 may continue to slide. The measured fracture strain is ~2.2−2.6% from PL data, and is ~2.9−3.5% from Raman measurement, which is smaller than the reported 6−11% breaking strain measured with nanoindentation on suspended circular MoS2 membrane49. We believe this is related to the fact that the MoS2 structure has two unclamped free edges, which could have dangling bonds and defects due to the mechanical exfoliation process. During bulging, cracks could initiate at the (mechanically) weak sites with defects, and then propagate to fracture the MoS2 structure. Study of the brittle fracture in freestanding graphene originating from cracks has been demonstrated,50 and we believe cracks in MoS2 will also lead to fracture between the two clamped edges instead of near the clamping area.

-12-

ACS Paragon Plus Environment

Displacement (Measurement) Displacement (Simulation)

1.0

εPDMS

9 8 7

0.8

6 5

0.6

4 0.4

3 2

0.2

1

(a)

Strain on PDMS εPDMS (%)

1.2

0.0

Silicone Clamp 31.03kPa

1L 3L

10µm

0 0

5

10

15

20

25

30

(b)

Differential Pressure (kPa)

εMoS , Raman- (%)

εMoS , PL (%)

2

2

1.90

1.67

2.21

2.61

PL

1.88 1.86 1.84 1.82 1.80

(c)

0

1

2

3

4

0 1.24 1.74 2.23 2.51

E12g Raman Shift (cm-1)

A Peak Energy (eV)

0 1.24 1.57

1L MoS2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Displacement, h (mm)

Page 13 of 24

5

6

7

8

9

3.54 E1+ 2g

Raman

E12g

386

384

382

(d)

εPDMS (%)

2.86

0

1

2

3

4

5

6

7

8

9

εPDMS(%)

Figure 3. Strain tuning of the MoS2 device shown in Figure 2, for the 1L region. (a) Calibration of the PDMS displacement and the corresponding strain in PDMS at varying differential gas pressures. (b) Optical image showing that the MoS2 finally fractures under applied strain, with the crack position indicated by the red dashed line and the red arrow. (c) & (d) Extracted (c) PL peak A and (d) Raman peak E12g position shift with applied strain in PDMS, with the corresponding strain in MoS2 indicated on the top axis (εMoS2, Raman- indicates the strain in MoS2 calculated from the strain response of E1-2g Raman peak). The shaded green area indicates that the device is still in the linear regime without sliding, while the pink region suggests that the MoS2 could have some sliding with substrate due to observed change in the slope. The MoS2 device finally fractures at strain of ~7-9.4% in PDMS, as shown by the red star.

-13-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 24

Similarly, for the 3L region in the MoS2 structure shown in Figure 2a and 2b, we summarize the PL peak position shifts with strain for both the direct transition peak A (Figure 4a) and indirect transition peak I (Figure 4b). The 3L MoS2 PL peaks show a similar tendency and position shift with strain as the 1L region, and the A peak position shifts at -30±3 meV/% strain, and I peak position shifts at -14±1 meV/% strain in the range of 0−1.2% strain (green area). Then the responsivity of peak position shift to strain changes at 1.2−7% strain (pink region). When the structure fractures, the I peak position shifts up to an even higher energy than that before applying strain. This is unexpected, and since the I peak position shifts toward the I peak position of 2L MoS2, it is possible that the 3L region that breaks has some MoS2 layer separation, so that it forms small regions with 2L MoS2. The fracture strain of MoS2 calculated from A and I PL peaks are quite similar, in the range of 1.9−2.2% strain, and are slightly lower than the fracture strain of the 1L region. Figure 4c shows that the Raman mode E12g of 3L also shows redshift with strain as the 1L region, with -0.5±0.1 cm-1/% strain. We do not observe Raman peak splitting for 3L region, and the fracture strain calculated from Raman measurement is 2.5-3.1% in MoS2.

Figure 4. Strain tuning of the MoS2 device shown in Figure 2, for the 3L region. (a) & (b) PL peak shift of (a) direct transition peak A and (b) indirect transition PL peak I, with applied strain in PDMS. (c) Raman peak E12g position shift with applied strain in PDMS. The extracted strain in MoS2 is indicated on the top axis, and the red star suggests fracturing of the MoS2 structure. -14-

ACS Paragon Plus Environment

Page 15 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Effect of Clamping on the Blown-Bubble Bulge Measurement.

We also determine the

effectiveness of our clamping approach by comparing the strain response of MoS2 structures with and without clamping, as shown in Figure 5. Figure 5a shows an optical image of a clamped sample containing a MoS2 structure with varying thicknesses. We perform blownbubble bulge measurement on the 1L and 2L regions under varying pressures, and the results are shown in Figure 5c and 5d, respectively. A similar redshift occurs for the A and I peaks in the PL spectra over a differential pressure range of 0 to 4.14 kPa (0.8% strain in PDMS), and no obvious sliding was observed at 2.07 kPa (0.25% strain in PDMS) with the silicone clamp (Figure 5b). The PL peak position shift with strain is summarized in Figure 5e (1L region) and 5f (2L region), showing that for the 1L region, A peak shifts with responsivity of -18.5 meV/% strain; and for the 2L region, A peak responsivity is -27.3 meV/% strain, and I peak responsivity is -89.1 meV/% strain. The A peak responsivity for the 1L region is smaller than the value measured earlier, probably because the 1L region is connected to the thicker MoS2 region, and the clamping condition is not ideal, which again proves the importance of clamping for the correct measurement of the strain response. We assume that for the amount of strain applied, the strain in PDMS is the same as that in MoS2 because the peak shift is still linear and the slope does not change. For another 1L MoS2 structure on PDMS without clamping (Figure 5g), we observe that MoS2 already detaches from the PDMS substrate and forms wrinkles at a differential pressure of 2.07 kPa (Figure 5h). From the PL spectra in Figure 5i and summary in Figure 5j, little peak shift of -2.9 meV/% strain is observed. The strain in MoS2 is hard to determine here because the MoS2 already delaminates.

The results show that the robust

clamping condition is essential in order to observe the correct shift in phonon modes and band structures till the fracture strain.

-15-

ACS Paragon Plus Environment

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 24

Figure 5. Comparison of MoS2 devices with and without silicone clamp. (a) & (b) Optical images of the clamped device (a) before and (b) under straining. (c) & (d) The PL peak shifts under applied pressure for (c) 1L region and (d) 2L region of the clamped device. (e) & (f) Summary of PL peak shifts of the clamped MoS2, for (e) A peak of the 1L region, and (f) I peak for the 2L region. (g) & (h) Optical images of the unclamped device (g) before and (h) under straining. (i) PL spectra under strain for the unclamped device, showing little peak shift. (j) Summary of the PL peak shift for the MoS2 structure without clamp.

-16-

ACS Paragon Plus Environment

Page 17 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

In conclusion, we demonstrate a blown-bubble bulge measurement technique and show that it can efficiently attain large strains in single- and few-layer MoS2 on top of flexible PDMS substrates, until the MoS2 crystals fracture. PL and Raman spectroscopy show that the bandgap and phonon modes shift with the applied strain, with the signatory peaks continuously redshifting upon increasing strain, up to the fracturing strain. The strain is found to be isotropic biaxial at