Atypical Defect-Mediated Photoluminescence and Resonance Raman

Subscriber access provided by TULANE UNIVERSITY

C: Physical Processes in Nanomaterials and Nanostructures

Atypical Defect-Mediated Photoluminescence and Resonance Raman Spectroscopy of Monolayer WS2 Jiake Li, Weitao Su, Fei Chen, Li Fu, Su Ding, Kaixin Song, Xiwei Huang, and Lijie Zhang J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b11647 • Publication Date (Web): 27 Jan 2019 Downloaded from http://pubs.acs.org on February 4, 2019

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 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 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.

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 25 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

The Journal of Physical Chemistry

Atypical Defect-Mediated Photoluminescence and Resonance Raman Spectroscopy of Monolayer WS2 1,*

1

1

1

2

2

Jiake Li, Weitao Su , Fei Chen , Li Fu , Su Ding , Kaixin Song , Xiwei Huang , Lijie Zhang

3,*

1

College of Materials and Environmental Engineering, Hangzhou Dianzi University, 310018, Hangzhou, China

2

College of Electronics and Information, Hangzhou Dianzi University, 310018, Hangzhou, China

3

College of Chemistry and Materials Engineering, Wenzhou University, 315201, Wenzhou, China.

Corresponding author: [email protected];

[email protected]

Abstract: Defects play an indispensable role in tuning the optical properties of two-dimensional(2D) materials. Herein, we study the influence of defects on the photoluminescence (PL) and resonance Raman spectra of as-grown monolayer(1L) WS2. Increasing the density of defects significantly lowers the excitonic binding energy by up to 110 meV. These defect-modified excitonic binding energies in1L-WS2 strongly mediate the Raman resonance condition, resulting in unexpected Raman intensity variations in the LA(M), 2LA(M) and

phonon modes. The sample

with the highest density of defects exhibits an almost temperature independent resonance in different Raman modes at low temperature, while the samples with low densities of defects exhibit clear resonance with decreasing temperature. This study will further increase our understanding of the role of defects in resonance Raman spectroscopy and of the phonon-exciton interaction in 1L-WS2.

Introduction Recently, 1L-WS2 has attracted great interest as an important 2D semiconductor due to its wide direct-energy gap and related unique electronic and optical properties1-3. WS2 has a periodic layered stacked structure, in which each period is composed of a plane of tungsten atoms sandwiched between two planes of sulfur atoms3-4. When the flake thickness of WS2 is reduced from a thick layer to a monolayer, its band structure transitions from an indirect band (with a band gap of ~1.4 eV) to a direct band with a much larger band gap of >2 eV3, 5. In addition to 1

ACS Paragon Plus Environment

The Journal of Physical Chemistry 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 2 of 25

the large band gap, significant spin-orbit coupling in WS2 results in a valance band degeneracy at the K point of the first Brillouin zone and a large valance band splitting of hundreds of meV6. In contrast to classical semiconductors, the absorption and PL of WS2 are dominated by exciton transitions7-8. The prominent absorption peaks centered at ~2.0, ~2.4 and ~2.8 eV of 1L-WS2 can be assigned to the A, B and C excitons, respectively7, where the A and B excitons are correlated to the spin-orbit splitting at the K point, while C has been assigned to a high energy degenerate transition at the Γ point7. These electronic and optical properties show the promise of 1L-WS2 in a wide range of applications, including valleytronics9, photosensors10-12, nanocavity lasers13, and light-emitting diodes (LED)14-15. The point group of 1L-WS2 is

16

. In the first Brillouin zone, the three

(the in-plane mode),

Raman-active zone-center phonons can be labeled (absent in conventional backscattering)16 and

17

(the out-of-plane mode) . In

addition, several first-order zone edge(M point) phonon modes can also be observed for samples with high densities of defects18-19. If the energy of the incident laser photon is far from the excitonic transition energy and the resonance of the exciton with the photon is negligible, the typically observed Raman bands of WS2 are the first-order Raman peaks of the zone center phonons, i.e.,

15, 17

and

. In

contrast, if the energy of the incident laser photon matches the excitonic energy, the Raman peaks of second-order phonons, mainly the overtones or combinations of zone center modes and the zone edge(M point), become dominant16, 18-20. Recently, Corro et al. thoroughly studied the resonance Raman spectroscopy of mechanically exfoliated thick and 1L-WS216. They found that the intensities of the

,

and 2LA(M) modes are greatly enhanced when the photon energies of the incident lasers match the energies of the A or B excitons. The resonance of second-order Raman modes has also been reported for 1L-MoS221, 1L-MoSe222, and 1L-WSe216. The resonances of Raman-active phonons provide much information about the electronic properties of 2D transition metal dichalcogenides (TMDs). Although a handful of works had reported the resonance Raman of mechanically exfoliated or CVD-grown 1L-WS2, the influence of intrinsic defects on 2


Page 3 of 25 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


the resonance Raman spectra of vapor phase deposited 1L-WS2 has been rarely studied. We noticed that McCreary et al. gave an initial study of the defect-related Raman spectroscopy of WS223. However, the detailed mechanism of the intrinsic defects and the related effect of low temperature on the resonance Raman spectroscopy are still unclear. In this study, we investigated the influence of intrinsic defects on the excitonic energy and resonance Raman spectra of 1L-WS2 deposited using the vapor transport (VT) method. Using room and low temperature PL measurements, we show that defects strongly modulate the excitonic energies. The samples with defects exhibit remarkably different defect density dependent resonance of the first- and second-order Raman-active phonon modes. This defect-mediated Raman resonance is much clearer at low temperatures, below 80K.

Methods 1L-WS2 Synthesis: Triangular WS2 monolayers were deposited on Si/SiO2(300 nm) substrates using a VT method(Figure1(a)). A quartz boat containing ten milligrams of WS2 powder(Aladdin, 99.9%) was placed in the center heating zone of a tube furnace(diameter: 1 inch), while growth substrates were placed ~18 cm away from the source in the downstream direction. The substrates were cleaned ultrasonically in ethanol for 10 minutes, followed by cleaning in hydrogen peroxide for 10 minutes, finally followed by 10 minutes of ultrasonication in deionized water. Ar/H2(5%) mixed gas was used as the carrier gas. Prior to the growth of WS2, the quartz tube was pumped down to ~1 Pa and then flushed with Ar/H2 for 3 minutes to remove oxygen and other impurity gases. We prepared four kinds of typical samples (T1-T4) with different densities of defects by changing the deposition parameters. For T1, the furnace was heated to 1000 °C over 50 minutes at a rate of 19.6 °C/min and held for 66 minutes before cooled down naturally to room temperature. The pressure was maintained at 200 Torr while the Ar/H2 flow rate was set to 13 sccm during the growth. For T2, a special holder was used to raise the substrate height inside the quartz tube. Other conditions were similar to those for T1. For T3, the growth 3



conditions were similar to those used for T1 except that the holding time was 65 minutes and the gas flow rate was 11 sccm. For T4, the growth conditions were similar to those used for T1, except that the pressure was 210 Torr and the rate of temperature increase was 22.2 °C/min. See Figure S1 in Section §1 of the ESI for the details and discussions of the growth process. Characterization: Optical images of WS2 were obtained using an optical microscope(MV3000, NOVEL). Micro Raman and PL measurements excited at laser wavelengths of 532 nm and 632.8 nm were conducted on a home-built Raman/PL system, consisting of an inverted microscope (Ti eclipse, Nikon) and a Raman spectrometer (iHR320, Horiba) attached to a CCD detector (Syncerity, Horiba). Other Raman spectra were measured using either a Labram(Horiba, laser wavelength 514.4 nm) or a Renishaw invia(laser wavelengths 325 and 488 nm) Raman spectrometer. The low temperature PL and Raman spectra were measured on a Renishaw invia Raman spectrometer with excitation wavelengths of 488 nm and 532 nm. A Linkam THMS-600 heating and freezing stage was used to obtain low temperature up to 80K. All the Raman spectra were calibrated using the Si Raman peak at 520 cm-1. The topographic images of samples were measured using an atomic force microscope (AFM, Innova, Bruker). Auger spectra measurements were conducted using nano-Auger electron spectroscopy.

Results and discussion The optical images of the WS2 samples T1-T3 are shown in Figure S2(a)-(c) of ESI. The optical image of the sample T4 is shown in Figure 1(b). All the deposited WS2 flakes exhibit isolated or merged triangular shapes. Their corresponding lateral sizes vary from several microns to ~30 μm. The uniform light purple contrast indicates that there are no thick layers within these flakes. The AFM topographic images of samples T1-T4 are shown in Figure S2(d)-(f) of ESI and Figure 1(c), respectively. A large number of nanocrystals can be seen on the surface of these four samples, which is a common feature in VT deposited WS224 and WSe225. Line profiles across the edges 4


Page 4 of 25

Page 5 of 25 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


along the marked positions are shown as insets in the corresponding topographic images. The heights of the 1L samples T1-T4 vary in a range of 0.65-1.0 nm, which is consistent with the typical thickness of 1L-WS2 (0.8 nm)20 considering the high surface roughness and accuracy of our AFM instrument. The PL spectra of the four samples T1-T4 excited using a 532 nm laser are shown in Figure 2(a). According to previous reports7, 26, only 1L-WS2 gives a single PL band, while thick WS2 with 2 to 5 layers gives a double wavelet PL band. It can be observed that each sample in Figure 2(a) shows a single PL band, indicating that these four samples are all monolayers. Moreover, three other major features can also be seen in Figure 2(a): (I)The full widths at half-maximum (FWHM) of the PL peaks of samples T1-T4 were calculated to be 69, 120, 142 and 160 meV, which are much wider than the reported value for 1L-WS2 (~50 meV)3; (II) the PL energies (EA) of samples T1-T4 gradually redshift from 1.987 to 1.873 eV, which are much smaller than those are typically observed for CVD-grown monolayer samples(2-2.1 eV)3,

19

; (III) the PL

intensity gradually decreases from sample T1 to T4. The PL intensity of sample T4 is only 1/6 that of sample T1. The PL peak of 1L-WS2 is composed of the radiation of the neutral exciton(A0, at the high energy side of the PL peak), the defect-bound exciton (XD, at the low energy side of the PL peak, representing the recombination of an electron trapped on a lattice defect with a valence band hole)27-29 and the charged exciton (A-, between A0 and XD)26,

30

. Following this assignment, the PL spectra in Figure 2(a) can be

deconvoluted well into A0, A-, and XD excitons(shown in Figure S(3) of ESI ). In Figure 2(a), the significant low-energy tails observed in the PL spectra of samples T1-T4 can be attributed to the PL emission of XD excitons. Figure 2(b) shows the fitted peak positions of these A0, A- and XD excitons for samples T1-T4. From sample T1 to T4, the photon energies of all these excitons gradually decrease: A0 decreases from 1.994 eV to 1.899 eV, A- decreases from 1.979 eV to 1.861 eV, and XD decreases from 1.910 eV to 1.759 eV. The energy differences between A0 and XD increase from 84meV for T1 to 140 meV for T4, which results in the PL peak width broadening in Figure 2(a). The reduced energies of the A0, A- and XD excitons further induce the redshift of the PL 5



peaks from sample T1 to T4 in Figure 2(a). The area ratio of the XD exciton in the whole PL envelope area is also plotted in Figure 2(b). It can be seen that the area ratio of XD gradually increases as a function of the decreasing excitonic energies. We noticed that the observed PL peak redshifts, width broadening and intensity quenching in samples T1-T4 are similar to those of ion-bombarded 1L-WS2 and 1L-MoS2 samples with increasing defects densities31-32. These difference in the PL are also similar to that of 1L-WS2 with defects prepared by CVD23. These similarities indicate increasing defect densities from sample T1 to T4. A density functional theory(DFT) calculation was carried out on perfect 1L-WS2 and 1L-WS2 with a single sulfur adatom(Sad) and a single sulfur vacancy(Vs), which are the most typical defects33(Figure S4 of ESI). The calculated band gap of perfect 1L-WS2 is 1.88 eV, which decreases to 1.83 and 1.76 eV for 1L-WS2 with Sad and Vs, respectively. Although the band gap values are underestimated by 0.2-0.3 eV, it is still indicative that the band gap of 1L-WS2 can be reduced by Sad or Vs, resulting in a redshift of the excitonic energy. This has also been seen in MoS234-35 and WSe233. Previous DFT studies have shown that the excitonic binding energies of monolayer WSe2 are only slightly changed, by ~0.01 eV, with different defects33, which is expected to be applicable to 1L-WS2. Because the excitonic transition energy is the sum of the observed excitonic PL energy and the excitonic binding energy1, 8, the decreasing excitonic energies of T1-T4 indicate a decreasing band gap from T1 to T4. In addition, Vs generates midgap states 0.53 eV below the minimum of the conduction band (Figure S4). These midgap states trap a large number of excitons and force them to nonradiatively transit to the ground state, resulting in a reduced PL intensity compared to that of the monolayer

sample with a low density of

defects36. Low temperature PL spectra for samples T1 and T4 below 80K are shown in Figure 2(c) and 2(d), respectively. In Figure 2(c)-(d), a clear splitting of the prominent PL peak can be seen. The PL spectra of sample T1 were further deconvoluted using Gaussian functions representing the A-, A0 and XD excitons, shown in Figure S5(a). For sample T4, two XD exciton bands, XDI and XDII, need to be considered to obtain good 6


Page 6 of 25

Page 7 of 25 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


fitting results(shown in Figure S5(b)). The PL spectra of sample T1 (Figure 2(c)) are very similar to that reported by Carozo et al. 34. The high energy mode can be assigned to the A0 exciton, while the low energy mode can be assigned to the Aexciton30. The shoulder at ~1.9 eV, observed when the temperature is lower than 120K, can be assigned to the XD exciton. The low temperature PL spectra of T4 (Figure2(d)) are quite similar to those of aged 1L-WS2 samples with high densities of adsorbates30. With decreasing temperature, the PL intensities of A0 and A- are greatly reduced, while the intensities of the XDI and XDII excitons become stronger. With decreasing temperature, the decreased intensities of free excitons(A0 and A-) but enhanced intensity of XD can be attributed to the redistribution of the exciton population due to the presence of defects37-39. At low temperatures

Atypical Defect-Mediated Photoluminescence and Resonance Raman

Recommend Documents