Thickness-dependent excitonic properties of atomically thin 2H-MoTe2
Li Jin-Huan1, Bing Dan1, Wu Zhang-Ting3, Wu Guo-Qing1, Bai Jing1, Du Ru-Xia1, Qi Zheng-Qing2, †
Department of Basic Teaching, Nanjing Tech University Pujiang Institute, Nanjing 211134, China
School of Network and Communication Engineering, Jinling Institute of Technology, Nanjing 211169, China
Lab for Nanoelectronics and NanoDevices, Department of Electronics Science and Technology, Hangzhou Dianzi University, Hangzhou 310018, China

 

† Corresponding author. E-mail: qi.zq@163.com

Project supported by the Natural Science Research Projects in Colleges and Universities of Jiangsu Province, China (Grant No. 18KJD140003).

Abstract

Two-dimensional (2D) 2H-MoTe2 is a promising semiconductor because of its small bandgap, strong absorption, and low thermal conductivity. In this paper, we systematically study the optical and excitonic properties of atomically thin 2H-MoTe2 (1–5 layers). Due to the fact that the optical contrast and Raman spectra of 2H-MoTe2 with different thicknesses exhibit distinctly different behaviors, we establish a quantitative method by using optical images and Raman spectra to directly identify the layers of 2H-MoTe2 thin films. Besides, excitonic states and binding energy in monolayer/bilayer 2H-MoTe2 are measured by temperature-dependent photoluminescence (PL) spectroscopy. At temperature T = 3.3 K, we can observe an exciton emission at ∼ 1.19 eV and trion emission at ∼ 1.16 eV for monolayer 2H-MoTe2. While at room temperature, the exciton emission and trion emission both disappear for their small binding energy. We determine the exciton binding energy to be 185 meV (179 meV), trion binding energy to be 20 meV (18 meV) for the monolayer (bilayer) 2H-MoTe2. The thoroughly studies of the excitonic states in atomically thin 2H-MoTe2 will provide guidance for future practical applications.

1. Introduction

Atomically thin transition metal dichalcogenides (TMDs) have recently attracted great attention for their special electronic and optical properties.[13] The atoms of TMDs are held together by strong covalent in-plane bonds but are oriented in the out-of-plane direction by weak van der Waals forces. Thus, few layered two-dimensional (2D) materials can be easily exfoliated from bulk crystal materials.[4,5] When these materials are thinned down to a single layer, the indirect band gap will transform into direct band gap,[68] this transformation not only enhances the quantum yield of photoluminescence (PL),[9] but also is a necessary condition for the observation of Valley Hall effect.[1012] In some TMDs, the indirect–direct bandgap transition between bulk and single layer provides a variety of optoelectronic applications ranging from photodetectors to light emitters.[13] The thickness of 2D materials strongly affects their optical, electronic, and other properties. In order to investigate the properties of 2D materials, it is necessary to identify the thickness quickly.

2H-MoTe2 is a new material, its layer-dependent bandgap ranges from 0.8 eV (bulk) to 1.1 eV (monolayer) as the thickness decreases and it exhibits direct-to-indirect bandgap transition.[14] The bandgap feature of 2H-MoTe2 flake suggests that it is favorable for visible and SWIR photodetection.[1517] Other distinctive properties have also been addressed in 2H-MoTe2. Its mobility at room temperature (RT), for example, can theoretically reach up to 200 cm2·V−1·s−1.[18] The gate voltage of several volts can drive phase transition from semiconductor to semimetal in monolayer 2H-MoTe2 by choosing appropriate dielectrics. With its unique properties, 2H-MoTe2 has a promising application in the fields of optoelectronics, energy storage, memristor, chemical and biological sensing, etc.[1921] In recent years, the optical properties of 2H-MoTe2 have been studied.[22] For example, the optical contrast,[23,24] PL and Raman spectra[25,26] have been observed. However, a systematically study in using these thickness-dependent properties to identify the layer number of 2H-MoTe2 is lacking. Moreover, owing to the spatial confinement and reduced dielectric screening, noticeable exciton effects are widely expected in monolayer crystals of TMDs.[27] Numerous experimental and theoretical investigations of neutral excitons and charged excitons (trions) have been conducted in 2D TMD semiconductors.[28,29] Most of these investigations focused on atomically thin MX2 (X = Se, S; M = W, Mo) 2D semiconductors. However, researches of the exciton binding energy in atomically thin 2H-MoTe2, which are of great significance for developing the novel 2H-MoTe2 based excitonic devices, are still lacking.

In this work, we systematically analyze the optical properties of mechanically exfoliated atomically thin 2H-MoTe2 by using atomic force microscope (AFM), optical contrast, and Raman spectroscopy. We identify the thickness of 1L–5L 2H-MoTe2 flake. Temperature-dependent PL measurement is used to determine the optical band gap and to probe excitonic states in few-layer 2H-MoTe2. From the PL spectra, we find that 2H-MoTe2 possesses a direct optical band gap of 1.19 eV. Therefore, 2H-MoTe2 is a new direct band gap 2D material, and its band gap is significantly lower than those of other TMDs available heretofore. More importantly, we obtain the exciton (trion) binding energy in monolayer/bilayer 2H-MoTe2.

2. Sample preparation

We used mechanical exfoliation to obtain atomically thin 2H-MoTe2 from bulk crystals (SPI Supplies) and transferred it into an Si wafer capped with a 300-nm-thick SiO2 layer. The exfoliation process was carried out in ambient air at room temperature. After exfoliation, tape residue was removed by being soaked in acetone, then rinsed with isopropanol and blow-dried with N2. The AFM was utilized to identify the thickness of 2H-MoTe2 as shown in Fig. 1. Figures 1(a) and 1(b) show the optical images of 2H-MoTe2 sample and the AFM curves traced with the red dashed line. The AFM curve shows that the step height from substrate to monolayer is about 1 nm, while the step height between the adjacent layers is about 0.7 nm each layer, consistent with previous measurements.[30] The larger step height difference above is possibly caused by the reflection changes between the substrate and the sample and also the presence of several adsorbates on the substrate.[31,32]

Fig. 1. (a) and (b) Optical microscope image and AFM curve (along the red dashed lines) of 2H-MoTe2 flakes.
3. Results and discussion

To further optically illustrate the layer number, the R, G, and B channel images of the optical image (Fig. 2(a)) are obtained by Matlab software,[31] and shown in Figs. 2(b)2(d). It can be seen that the optical contrast of 2H-MoTe2 sample varies with thickness obviously in the R channel (Fig. 2(b)) and G channel (Fig. 2(c)). However, in the B channel (Fig. 2(d)), the optical contrast is hard to distinguish between 2H-MoTe2 sample and substrate. To quantify the optical contrast as shown in Fig. 2(e), the variation of the contrast value with layer thickness is measured with an optical microscope. The contrast value C is calculated from the equation C = (CsubCsam)/Csub. Here, Csub is the contrast value taken from the substrate, and Csam is the contrast value taken from the 2H-MoTe2 sample.[33] The contrast values of R channel are 0.22 (1L), 0.34 (2L), 0.42 (3L), 0.45 (4L), and 0.39 (5L), respectively. For 1–3 layers of 2H-MoTe2, the contrast value of R channel increases linearly. The contrast of G channel decreases linearly with 3–5 layers. For 2H-MoTe2 samples with different layers, the change of contrast value of B channel is not obvious, so the number of layers of 2H-MoTe2 can be identified by the contrast between R channel and G channel.

Fig. 2. (a) Optical image of 2H-MoTe2 sample obtained with mechanical exfoliation; (b)–(d) R, G, and B channel images of 2H-MoTe2 sample; (e) optical contrast values of 2H-MoTe2 samples with different thicknesses.

Raman spectroscopy as a noninvasive method can also be used to measure the thickness of 2D material.[34] Figure 3(a) shows three typical Raman modes of 2H-MoTe2: the out-of-plane A1g mode (∼ 170 cm−1), the in-plane mode (∼ 233 cm−1), and the phonon mode (∼ 290 cm−1).[24] The position of the A1g peak and the peak upshift with thickness decreasing (Fig. 3(b)), the change in the peak position is due to the effect of boundary surface and interlayer interaction with thickness decreasing. Furthermore, a strong peak at ∼ 290 cm−1 emerges in the atomically thin 2H-MoTe2. This peak is not observed in the bulk nor monolayer 2H-MoTe2 crystal, but the intensity is very strong in bilayer and decreases with thickness increasing. In Fig. 3(c), it seems that the A1g mode of 3L–5L 2H-MoTe2 exhibits multiple peaks, which is well known as Davydov splitting in bulk and bilayer TMDs.[35,36] The generation of Dovydov splitting can be attributed to the interaction between layers. In addition, with the increase of layers’ number, the A1g peak is blue-shifted. According to Fig. 3(d), the intensity ratios of for 2–5 layers are 0.37, 0.22, 0.19, and 0.15, respectively, their changes with layer are obvious. Therefore, the layers of 2H-MoTe2 can also be determined by the intensity ratio of . The dependence of the peak position of mode on thickness is also studied as shown in the inset, it is found that the mode is shifted up by 1.5 cm−1, with thickness decreasing from five layers to a single layer.

Fig. 3. (a) Three Raman-active vibration modes of layered 2H-MoTe2, , A1g, and respectively, (b) Raman spectra of 1L–5L and bulk 2H-MoTe2, (c) A1g modes in 1L–5L 2H-MoTe2, (d) intensity ratio of versus number of layers, with inset showing the peak position of the mode as a function of thickness.

Figures 4(a) and 4(b) display the photoluminescence (PL) spectra of monolayer and bilayer 2H-MoTe2 varying with temperature, respectively. The power and wavelength of incident laser are, respecectively, 100 μW and 532 nm for Raman and PL measurement. At 3.3 K, the 1L 2H-MoTe2 exhibits two PL peaks at about 1.16 eV and 1.19 eV, which have been studied and assigned to the emission from trion and exciton, respectively.[37] At room temperature, we observed no PL of 2H-MoTe2, which is mainly because of its small binding energy, vulnerable to external environment (such as tem perature). With the increase of temperature, electrons or holes change from exciton radiation recombination to non-radiation recombination due to thermal disturbance, the intensities of neutral and charged exciton peaks appear at low temperature gradually decrease with temperature increasing, and almost disappear at 150 K. When the temperature exceeds 60 K, it mainly comes from neutral exciton luminescence, for the binding energy of neutral exciton is larger than that of charged exciton.[29] From the difference in energy between the exciton peak and trion peak, we can calculate the trion binding energy (Figs. 4(c) and 4(d)). The binding energy values of the monolayer and bilayer charged exciton of 2H-MoTe2 are 21 meV and 19 meV, respectively. The binding energy of charged exciton in other ultra-thin 2D semiconductors have also been reported, such as 30 meV for monolayer MoSe2, 18 meV for monolayer MoS2, and 30 meV for bilayer WSe2.[3739] In addition, with the increase of temperature, all peaks of 2H-MoTe2 in monolayer and bilayer are red-shifted, mainly because the band gap decreases with the increase of temperature. The relationship between band gap and temperature is as follows:[40] Eg (T) = Eg (0) − Sħω〉[coth (〈ħω〉/2kBT)]. Here E(0) is the bandgap at T = 0 K, S is the dimensionless constant, 〈ħω〉 represents the average acoustic phonon energy involved in the electron-phonon interaction, and kB is the Boltzmann constant. As the temperature increases, Eg(T) decreases gradually, so the PL peak exhibits the red shift. Exciton–phonon interaction has an important influence on the movement of exciton peak. When the exciton moves in the lattice, it will interact with the phonon through the scattering. With the increase of temperature, the scattering increases. Therefore, the exciton energy changes due to the exciton – phonon interaction.

Fig. 4. PL spectra of (a) 1L and (b) 2L 2H-MoTe2 versus temperature, exciton and trion peak positions of mono (c) and bilayer (d) 2H-MoTe2 as a function of temperature.

Figure 4 shows that the PL intensity of neutral and charged exciton decrease with temperature increasing. In the case of steady state, we can consider that the change of exciton intensity with temperature is mainly a reflection of the exciton population, and the excitons are in complete thermal equilibrium. The process of thermal dissociation can be used to describe the change of exciton population (N) with temperature T :[41] N (T) = N0/(1 + (τ/τ0)eEA/kBT) where N0 is the population of excitons at 0 K (i.e., peak intensity), τ is the lifetime of excitons, τ0 is the effective scattering time, and EA is the activation energy. E describes the necessary energy for the thermal quenching of exciton luminescence. Fitting the data of PL peak intensity versus temperature (Figs. 4(a) and 4(b)) with this equation, we obtain E = 185 meV and 20 meV (179 meV and 18 meV) for the exciton and trion in monolayer (bilayer) (Fig. 5), respectively. This result is reasonable compared with that from similar TMDs like monolayer WSe2 (210 meV).[42]

Fig. 5. Plots of exciton and trion intensities versus temperature, where solid lines are fits to data by using thermal dissociation model.
4. Conclusions

In this work, we study the optical properties of atomically thin 2H-MoTe2 to identify the layers directly. This identification utilizes the layer-dependence of the optical contrast, and Raman spectra of 2H-MoTe2 will provide a fast, nondestructive, easy-to-use, and accurate method to identify the number of 2H-MoTe2 layers. By analyzing the temperature-dependent PL spectra, the excitonic states and exciton binding energy values of single-layer and double-layer) 2H-MoTe2 are investigated. The exploration of the excitonic state in 2H-MoTe2 is an important step towards materials optimization for the near-infrared photodetection.

Reference
[1] Guo X T Wang W H Nan H Y Yu Y F Jiang J Zhao W W Li J H Zafar Z Xiang N Ni Z H Hu W D You Y M Ni Z H 2016 Optica 3 1066
[2] Du R X Wang W H Du J X Guo X T Liu E Bing D Bai J 2016 Appl. Phys. Express 9 115101
[3] Zhang H Cheng H M Ye P D 2018 Chem. Soc. Rev. 47 6009
[4] Lv R T Robinson J Schaak R Sun D Sun Y F Mallouk T Terrones M 2014 Acc. Chem. Res. 48 56
[5] Splendiani A Sun L Zhang Y Li T Kim J Chim C Y Galli G Wang F 2010 Nano Lett. 10 1271
[6] Zhao W Ghorannevis Z Chu L Toh M Kloc C Tan P H Eda G 2012 ACS Nano 7 791
[7] Tongay S Zhou J Ataca C Lo K Matthews T S Li J Grossman J C Wu J 2012 Nano Lett. 12 5576
[8] Zhang Y Chang T R Zhou B Cui Y T Yan H Liu Z Schmitt F Lee J Moore R Chen Y Lin H Jeng H T Mo S K Hussain Z Bansil A Shen Z X 2014 Nat. Nanotechnol. 9 111
[9] Lezama I G Arora A Ubaldini A Barreteau C Giannini E Potemski M Morpurgo A F 2015 I Nano Lett. 15 2336
[10] Cao T Wang G Han W Ye H Zhu C Shi J Niu Q Tan P Wang E Liu B Feng J 2012 Nat. Commun. 3 887
[11] Mak K F He K Shan J Heinz T F 2012 Nat. Nanotechnol. 7 494
[12] Zhang Y J Oka T Suzuki R Ye J T Iwasa Y 2014 Science 344 725
[13] Eda G Maier S A 2013 ACS Nano 7 5660
[14] Yang J T Myint Y W Pei J Macdonald D Zheng J C Lu Y 2015 ACS Nano 9 6603
[15] Huang H Wang J Hu W Liao L Wang P Wang X Gong F Chen Y Wu G Luo W Shen H Lin T Sun J Meng X Chen X Chu J 2016 Nanotechnology 27 445201
[16] Octon T Nagareddy V Russo S Craciun M Wright C 2016 Adv. Opt. Mater. 4 1750
[17] Yu W Li S Zhang Y Ma W Sun T Yuan J Fu K Bao Q 2017 Small 13 1700268
[18] Conan A Bonnet A Zoaeter M Ramoul D 1984 Phys. Status Solidi B 124 403
[19] Lin Y F Xu Y Wang S T Li S L Yamamoto M Aparecido-Ferreira A Li W Sun W H Nakaharai S Jian W B 2014 Adv. Mater. 26 3263
[20] Feng Z Xie Y Chen J Yu Y Zheng S Zhang R Pang W 2017 2D Materials 4 025018
[21] Sun L Ding M Li J Yang L Lou X Xie Z Chang H 2019 Appl. Surf. Sci. 496 143687
[22] Li Y Duerloo K A Wauson K Reed E J 2016 Nat. Commun. 7 10671
[23] Ruppert C Aslan O B Heinz T F 2014 Nano Lett. 14 6231
[24] Yamamoto M Wang S T Ni M Lin Y F Li S L Aikawa S Jian W B Ueno K Wakabayashi K Tsukagoshi K 2014 ACS Nano 8 3895
[25] Guo H H Yang T Yamamoto M Zhou L Ishikawa R Ueno K Tsukagoshi K Zhang Z D Dresselhaus M S Saito R 2015 Phys. Rev. B 91 205415
[26] Grzeszczyk M Gołasa K Zinkiewicz M Nogajewski K Molas M R Potemski M Wysmołek A Babiński A 2016 2D Material 3 025010
[27] Qiu D Y da Jornada F H Louie S G 2013 Phys. Rev. Lett. 111 216805
[28] Berkelbach T Hybertsen M Reichman D 2013 Phys. Rev. B 88 045318
[29] Zhu B Chen X Cui X 2015 Sci. Rep. 5 9218
[30] Chang Y M Lin C Y Lin Y F Tsukagoshi K 2016 Jpn. J Appl. Phys. 55 1102A1
[31] Nemes-Incze P Osvath Z Kamaras K Biro L P 2008 Carbon 46 1435
[32] Lee C Yan H Brus L E Heinz T F Hone J Ryu S 2010 ACS Nano 4 2695
[33] Wang Y Y Gao R X Ni Z H He H Guo S P Yang H P Cong C X Yu T 2012 Nanotechnology 23 495713
[34] Ni Z H Wang Y Y Yu T Shen Z X 2008 Nano Res. 1 273
[35] Song Q J Tan Q H Zhang X Wu J B Sheng B W Wan Y Wang X Q Dai L Tan P H 2016 Phys. Rev. B 93 115409
[36] Kim S Kim K Lee J U Cheong H 2017 2D Materials 4 045002
[37] Ross J S Wu S Yu H Ghimire N J Jones A M Aivazian G Yan J Mandrus D G Xiao D Yao W Xu X 2013 Nat. Commun. 4 1474
[38] Mak K F He K Lee C Lee G H Hone J Heinz T F Shan J 2013 Nat. Mater. 12 207
[39] Jones A M Yu H Ross J S Klement P Ghimire N J Yan J Q Mandrus D G Yao W Xu X D 2014 Nat. Phys. 10 130
[40] Huang J Hoang T B Mikkelsen M H 2016 Sci. Rep. 6 22414
[41] Kulyuk L Charron L Fortin E 2003 Phys. Rev. B 68 075314
[42] Zhang Y Ugeda M M Jin C H Shi S F Bradley A J Martín-Recio A Zhou B 2016 Nano Lett. 16 2485