Quick Optical Identification of the Defect Formation in Monolayer WSe2 for Growth Optimization
- 108 Downloads
Bottom-up epitaxy has been widely applied for transition metal dichalcogenides (TMDCs) growth. However, this method usually leads to a high density of defects in the crystal, which limits its optoelectronic performance. Here, we show the effect of growth temperature on the defect formation, optical performance, and crystal stability in monolayer WSe2 via a combination of Raman and photoluminescence (PL) spectroscopy study. We found that the defect formation and distribution in monolayer WSe2 are closely related to the growth temperature. These defect density and distribution can be controlled by adjusting the growth temperature. Aging experiments directly demonstrate that these defects are an active center for the decomposition process. Instead, monolayer WSe2 grown under optimal conditions shows a strong and uniform emission dominated by neutral exciton at room temperature. The results provide an effective approach to optimize TMDCs growth.
KeywordsWSe2 Defects Crystal stability Photoluminescence Raman scattering
Atomic force microscope
Chemical vapor deposition
Full width at half maximum
standard-state cubic centimeter per minute
Scanning electron microscope
Scanning tunneling microscopy
Transmission electron microscopy
Transition metal dichalcogenides
Ultrathin TMDCs (MX2, M = Mo, W; X = Se, S, etc.) have been widely applied in the photonic and optoelectronic application fields, such as photodetectors [1, 2, 3, 4], ultrathin transistors [5, 6], photovoltaic devices [7, 8], sensors [9, 10], and electrocatalysis . Compared with mechanical exfoliation method, chemical vapor deposition (CVD) shows great advantages in massive production, morphology, and structure controlling [12, 13, 14, 15], which are highly desired for large-area flexible material development and optoelectronic device applications [2, 16, 17, 18]. However, the formation of lattice defects in two-dimensional (2D) materials during the CVD growth is detrimental to its photoelectric properties, device performance, and even the crystal stability. For example, the hole mobility of WSe2 field-effect transistor fabricated using CVD grown monolayer is far below the theoretical predictions . The defect formation-induced nonuniform photoluminescence (PL) emission distribution has been widely observed in the grown TMDCs monolayer [20, 21, 22, 23, 24]. CVD-grown TMDCs monolayer shows poor lattice stability in the air . The high defect density in CVD-grown 2D materials significantly limits their device performance and stability, especially for devices exposed to the air for a long time.
The most direct and effective methods for 2D materials defect detection are transmission electron microscopy (TEM)  and scanning tunneling microscopy (STM) technique . But these methods usually require sample transferring which could cause new defects. In addition, these methods are time-consuming and only detect the defects in a small area. For the growth optimization, a quick and nondestructive evaluation method is highly demanded. Raman spectroscopy is an important and nondestructive method to probe the lattice vibration, lattice distortion, and electronic properties of materials [28, 29]. For instance, the XeF2 treatment-induced defects in WSe2 have been studied by comparing the E12g peak intensity, the peak shift, and the full width at half maximum (FWHM) . PL spectroscopy shows advantages in quickly determining the optical properties and detecting the electronic structure TMDCs without damaging. So it is widely used to study the optical properties of TMDCs [2, 31, 32]. In addition, PL is quite sensitive to the excitons, trions, and defects in monolayer TMDCs [33, 34, 35, 36]. Rosenberger et al. show an inverse relationship between the PL intensity of monolayer WS2 and defect density . Further research shows that the weak PL is mostly due to the formation of negatively charged excitons . Therefore, optical characterization offers a quick and nondestructive method to evaluate the localized defects and crystal quality of TMDCs.
Growth time and growth temperature are the two most important parameters affecting the growth of 2D materials. These effects on the growth duration of CVD-grown WSe2 monolayer have been reported before . Therefore, in this work, we try to focus on the optical property difference of WSe2 grown at different temperatures and study the defect-induced crystal stability differences. The optical performance and the lattice quality are examined using confocal Raman and PL techniques for growth optimization. The crystal defects are found to weaken the PL emission intensity and lead to a nonuniform emission distribution in the triangle WSe2 domain due to defect density difference. Moreover, these defects cause a low energy emission peak in the PL spectrum, as observed in both room temperature and low-temperature PL spectra. In addition to the negative effect on the optical performance, the defects deteriorate the crystal stability in the air, resulting in faster decomposition rate of WSe2. Based on the optical characterization results, we found that there exists an optical growth temperature for WSe2. In our case, this temperature is 920 °C. Either reducing or increasing the growth temperature impacts the optical properties and crystal stability of monolayer WSe2. These results provide an approach for us to optimize the optical properties and crystal stability of 2D materials .
Synthesis of Monolayer WSe2
Monolayer WSe2 was synthesized using high-purity Se powder (Alfa-Aesar 99.999%) and WO3 powder (Aladdin 99.99%) using a 2-inch-diameter quartz tube furnace. The Se powders (30 mg) were placed in a quartz boat at the first heating zone. WO3 powders (100 mg) were placed in a quartz boat at the second heating zone. The distance between the Se powder and WO3 powder is about 25 cm. c-plane (0001) sapphire substrates were cleaned and placed at downstream (5~10 cm) of the WO3 solid sources. Before the experiments, the chamber was pumped about 10 min and flushed with high-purity Ar carrier gas (99.9999 %) under a flow of 200 standard-state cubic centimeter per minute (sccm) at room temperature to remove the oxygen contamination. After that, 10% H2 and Ar mixture gas with a flow of 50 sccm was introduced into the furnace at an ambient pressure. The second heating zone was heated to the target temperature (860~940 °C) at a ramping rate of 20 °C/min. After that, the temperature was maintained at the growth temperature for 6 min. Meanwhile, the first heating zone was kept at 320 °C. After growth, the furnace was cooled to room temperature.
The morphology of as-grown WSe2 was examined using an optical microscopy (NPLANEPi100X). Raman scattering and micro-PL measurements were performed using a Renishaw system (inVia Qontor). The excitation was pumped through an objective lens (× 100) with a green (532 nm) laser and 1800 lines/mm grating. Atomic force microscope (AFM) measurements were performed using an Agilent system (Agilent 5500, Digital Instruments, tapping mode). The morphology changes of monolayer WSe2 were examined by scanning electron microscopy (SEM, TESCAN MIRA3 LMU).
Results and Discussion
The Raman and PL emission spectra from the center and edge of monolayer WSe2 grown at 900 °C are compared in Fig. 3. The obtained PL spectra from center position is deconvoluted into three peaks: neutral exciton at ~ 1.624 eV (marked as A) [51, 52], trion at 1.60 eV (marked as A+) [29, 52], and an unknown emission peak (marked as D) around 1.53 eV (the detailed fitting basis are shown in Additional file 1: Figures S6–S8). Figure 3b shows the PL emission is dominated by the A+ in the center position. The binding energy for A+ is estimated to be about 24 meV, which is the energy difference between trions and neutral exciton . It fits perfectly with the value of positive trion in the literature [33, 35], where the trion consists of two holes (h+) and an electron (e−). Indeed, recent studies reveal that CVD-grown WSe2 is usually p-type due to the formation of tungsten vacancy . These results are consistent with the general rules of doping effects in semiconductors. During the power-dependent PL experiments, D emission quickly saturates (see Additional file 1: Figure S7 in the SI), suggesting that the unknown emission is actually caused by the lattice defects, as observed in other reports [24, 33, 51, 52]. In comparison, the emission from the edge does not contain this defect-related peak. Instead, the emission peak is much narrower and stronger, consisting of mainly neutral exciton peak with trion peak as a shoulder. During the power-dependent PL experiments, the FWHM of WSe2 on both center and edge does not change with power, indicating no signs of local heating effect (see Additional file 1: Figure S8 in the SI) [51, 60]. This defect-related emission peak becomes more obvious at low temperature (77 K), as compared in Fig. 3d. The PL spectrum at 77 K from the center region consists of three emission peaks. Through calculations, the binding energies of monolayer WSe2 for trion (A+) and defect-related emission are around 24 meV and 100 meV, respectively, which are consistent with our room temperature PL fitting results.
These results confirm the existence of the crystal defect in the CVD-grown WSe2 monolayer. These defects are centers for nonradioactive recombination, thus dropping the photon emission efficiency [24, 61]. Moreover, the defect density is position and growth condition dependent, leading to different emission distribution pattern in Fig. 2. Under poor growth conditions, monolayer WSe2 can still form. However, a large proportion of area is highly defected and contains only a small area with high crystal purity. PL spectrum and mapping provide a quick method to evaluate its crystal quality and guide the growth optimization. According to the above analysis, the monolayer WSe2 growth at lower growth temperature shows a weaker crystal quality, which could be due to insufficient reaction between the WO3-x and Se gas [62, 63]. Improving the temperature could thus overcome the reaction barrier and form WSe2 with high crystal quality (920 °C). However, keeping increasing the temperature (940 °C) could lead to the decomposition of the formed monolayer WSe2 under insufficient Se gas protection . Thus, the defect formation mechanism could vary at different growth temperatures, thereby leading to different emission distribution patterns. We found that the PL intensity of inner region of the triangle is the lowest. The decrease of PL intensity suggests that the crystal defects of the WSe2 were produced from the center of the triangle, which is consistent with previous reports . In addition, the probability of lattice distortion along the armchair (see Fig. 2a) direction is larger for monolayer WSe2 at 900 °C. As the WSe2 grown from the center of the triangle to the three angle edges of the triangle, the crystal quality of WSe2 is getting better.
The PL emission in Fig. 4g shows a similar trend. Compared with the data measured 90 days before, the PL peak position and emission intensity of the center region are blue-shifted by ~ 60 meV and decreased 7 times, respectively. Moreover, the FWHM is broadened by ~ 17 meV. In contrast, the PL peak position and FWHM of the edge are nearly the same and the emission intensity only drops to half of the intensity measured at 90 days before. Using the same approach, we found that the crystal deterioration process in monolayer WSe2 grown at 940 °C shows the same mechanism: the higher the crystal quality, the slower is the decomposition.
In order to better understand the aging process, the morphology evolution of monolayer WSe2 grown at 900 °C with time is shown in Fig. 5. The aged region starts from the center of the triangle (see Fig. 5b). As the aging time increases, WSe2 decomposes gradually from the center to the vertex of the triangle as shown in Fig. 5c. After 180 days, WSe2 at the center of the triangle and the three angular positions have been substantially decomposed completely. At this time, the PL in the center and triangle has quenched. Raman scattering in these decomposed areas shows no signal of vibration mode of WSe2, confirming the complete decomposition of WSe2 crystal. The aging study of a single layer of WSe2 grown at 900 °C further demonstrates that the location of the decomposition agrees very well with our previously measured PL mapping results. According to the above discussions, the critical factor affecting the stability of WSe2 is the unwanted defect formation during the CVD growth. PL and Raman spectrum provides an easy approach to quickly examine the crystal quality to guide the growth optimization towards 2D layer with the purest crystal quality.
In summary, we study the role of growth temperature on crystal defect formation and crystal stability of monolayer WSe2 on a sapphire substrate. PL and Raman spectroscopy techniques are applied to quickly identify the crystal quality, stability, and defect distribution of as-grown monolayer WSe2 at different conditions. Through this characterization approach, the optimal growth temperature for monolayer WSe2 is obtained at 920 °C. Either reducing or increasing the growth temperature leads to the formation of a higher defect density. At lower growth temperature, the defect formation is probably due to the unfully decomposed WO3-x precursor. The defects start to form at the nucleus center and then proceed along the armchair direction of the crystal, forming an inner triangular shape with a high density of defects and lower PL emission intensity. Above the optimal growth temperature, the defect distribution shows another pattern and starts from the edge, probably due to the decomposition of WSe2 at such a high temperature. PL emission shows that photon emission in the defected region is dominated by trions while neutral exciton emission is prominent in the WSe2 monolayer with better crystal quality. The aging experiment further proved that the region with a higher defect density can easily combine with O and OH, deteriorating its lattice stability. These results offer insights into the optimum synthesis of various 2D materials and the potential applications in the field of optoelectronics.
We thank the Analytical and Testing of Hunan Key Laboratory of Super Micro-structure and Ultrafast Process, School of Physics and Electronics, Central South University, as well as School of Materials Science and Engineering, Central South University.
ST, XY, and LF conceived the idea of experiments. LF, LL, and JD carried out the experiments. ST, XY, HC, HH, GC, and JH participated in the discussion and analysis of the experimental result. LF and XY co-wrote this paper. ST, XY, and HC modified the manuscript. All authors read and approved the final manuscript.
This work is supported by the National Natural Science Foundation of China (No.51702368, No.11674401, and No.11804387), the Hunan Provincial Natural Science Foundation of China (2018JJ3684), the Innovation-Driven Project of Central South University (2018CX045), the Fundamental Research Funds for the Central South University (No. 1053320181264), the Independent Exploration and Innovation Project for Postgraduates of Central South University (2018ttzs103), and the Open-End Fund for the Valuable and Precision Instruments of Central South University (CSUZC201826).
The authors declare that they have no competing interests.
- 8.Zheng X, Wei Y, Liu J, Wang S, Shi J, Yang H, Peng G, Deng C, Luo W, Zhao Y, Li Y, Sun K, Wan W, Xie H, Gao Y, Zhang X, Huang H (2019) A homogeneous p-n junction diode by selective doping of few layer MoSe2 using ultraviolet ozone for high-performance photovoltaic devices. Nanoscale 11(28):13469–13476CrossRefGoogle Scholar
- 9.Medina H, Li JG, Su TY, Lan YW, Lee SH, Chen CW, Chen YZ, Manikandan A, Tsai SH, Navabi A, Zhu XD, Shih YC, Lin WS, Yang JH, Thomas SR, Wu BW, Shen CH, Shieh JM, Lin HN, Javey A, Wang KL, Chueh YL (2017) Wafer-scale growth of WSe2 monolayers toward phase-engineered hybrid WOx/WSe2 films with sub-ppb NOx gas sensing by a low-temperature plasma-assisted selenization process. Chem Mat 29(4):1587–1598CrossRefGoogle Scholar
- 10.Ai YF, Hsu TH, Wu DC, Lee L, Chen JH, Chen YZ, Wu SC, Wu C, Wang ZMM, Chueh YL (2018) An ultrasensitive flexible pressure sensor for multimodal wearable electronic skins based on large-scale polystyrene ball@reduced graphene-oxide core-shell nanoparticles. J Mater Chem C 6(20):5514–5520CrossRefGoogle Scholar
- 15.Wu D, Shi J, Zheng X, Liu J, Dou W, Gao Y, Yuan X, Ouyang F, Huang H (2019) CVD grown MoS2 nanoribbons on MoS2 covered sapphire (0001) without catalysts. Phys Status Solidi-Rapid Res Lett 13:1900063Google Scholar
- 40.Wang K, Huang B, Tian M, Ceballos F, Lin MW, Mahjouri-Samani M, Boulesbaa A, Puretzky AA, Rouleau CM, Yoon M, Zhao H, Xiao K, Duscher G, Geohegan DB (2016) Interlayer coupling in twisted WSe2/WS2 bilayer heterostructures revealed by optical spectroscopy. ACS Nano 10(7):6612–6622CrossRefGoogle Scholar
- 62.Zhou JD, Lin JH, Huang XW, Zhou Y, Chen Y, Xia J, Wang H, Xie Y, Yu HM, Lei JC, Wu D, Liu FC, Fu QD, Zeng QS, Hsu CH, Yang CL, Lu L, Yu T, Shen ZX, Lin H, Yakobson BI, Liu Q, Suenaga K, Liu GT, Liu Z (2018) A library of atomically thin metal chalcogenides. Nature 556(7701):355–359CrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.