Chinese Journal of Chemical Physics  2018, Vol. 31 Issue (6): 767-771

#### The article information

Ze Wang, Ji-hao Wang, Wei-feng Ge, Wen-jie Meng, Jing Zhang, Qi-yuan Feng, Yu-bin Hou, Qing-you Lu

Direct Observation of Transition Metal Dichalcogenides in Liquid with Scanning Tunneling Microscopy

Chinese Journal of Chemical Physics, 2018, 31(6): 767-771

http://dx.doi.org/10.1063/1674-0068/31/cjcp1803040

### Article history

Received on: March 16, 2018
Accepted on: April 27, 2018
Direct Observation of Transition Metal Dichalcogenides in Liquid with Scanning Tunneling Microscopy
Ze Wanga,b, Ji-hao Wangb, Wei-feng Gea, Wen-jie Mengb, Jing Zhangb, Qi-yuan Fenga, Yu-bin Houb, Qing-you Lua,b,c,d
Dated: Received on March 16, 2018; Accepted on April 27, 2018
a. Hefei National Laboratory for Physical Sciences at the Microscale, University of Science and Technology of China, Hefei 230026, China;
b. Anhui Province Key Laboratory of Condensed Matter Physics at Extreme Conditions, High Magnetic Field Laboratory, Chinese Academy of Sciences, Hefei 230026, China;
c. Hefei Science Center, Chinese Academy of Sciences, Hefei 230031, China;
d. Collaborative Innovation Centre of Advanced Microstructures, Nanjing University, Nanjing 210093, China
*Author to whom correspondence should be addressed. Ji-hao Wang, E-mail:jihaowang@hmfl.ac.cn; Qing-you Lu, E-mail:qxl@ustc.edu.cn
Abstract: We present atomic-resolution images of TiSe$_2$, MoTe$_2$ and TaS$_2$ single crystals in liquid condition using our home-built scanning tunneling microscopy (STM). By facilely cleaving of single crystals in liquid, we were able to keep the fresh surface not oxidized within a few hours. Using the high-stable home-built STM, we have obtained atomic resolution images of TiSe$_2$ accompanied with the single atom defects as well as the triangle defects in solution for the first time. Besides, the superstructure of MoTe$_2$ and hexagonal charge-density wave domain structure in nearly commensurate phase of TaS$_2$ were also obtained at room temperature (295 K). Our results provide a more efficient method in investigating the lively surface of transition metal dichalcogenides. Besides, the high stable liquid-phase STM will support the further investigations in liquid-phase catalysis or electrochemistry.
Key words: Scanning tunneling microscopy    Liquid    Defect    Charge-density wave
Ⅰ. INTRODUCTION

Transition metal dichalcogenides (TMDS) are layered compounds consisting of a layer of transition metal (such as titanium, niobium, molybdenum etc.) sandwiched between two layers of chalcogen (such as sulfur, selenium, etc.) to form chalcogen-metal-chalcogen sandwiches, the sandwich structure stacks together via weak van der Waals forces to form a single crystal. Owing to their multiple interesting physical properties, these compounds have attracted considerable attentions in condensed matter physics and materials science [1]. TiSe$_2$, a member of TMDS, is a two-dimensional layered material with a charge density wave (CDW) transition temperature of $T_{\textrm{CDW}}$$\approx200 K. Although the band structure of TiSe_2 has already been determined by angle-resolved photoemission spectroscopy [2], it is still uncertain whether TiSe_2 is a semiconductor [3] or a semimetal [4]. The structure of TaS_2 is similar to that of TiSe_2, while the CDW structure in NC phase of TaS_2 can be observed at room temperature [5, 6]. Besides, the CDW structure of TaS_2 is easily affected by the probe tip. MoTe_2, Te-based TMDS, was studied only in recent years. The focus of earlier investigations has been on thermoelectric phenomena [7] and bulk transport [8], but reports on lattice structure and defect structure of MoTe_2 are very rare. The scanning tunneling microscopy (STM) can be used to investigate the atomic structure and defects on the surface of crystals as it probes the local density state of the fermi energy. Compared with TEM, electron diffraction, and X-ray diffraction (XRD) etc., STM has the ability of direct observation on samples surface. This method has been proven to be quite efficient on imaging materials with layered structure, especially for the studies on graphite [9, 10]. Different from graphite, surface properties of TiSe_2 are much lively, tending to oxidization in air. Although multiple STM images of TiSe_2 have been reported, its availability was only limited to the ultra high vacuum (UHV) condition [11], STM image of TiSe_2 with atomic resolution in air was never seen. Compared with TiSe_2, TaS_2 is a little more stable in air. When exposed in air for long time, the fresh surface of TaS_2 will also oxidize too severely to give an atomic resolution STM image. There are a few STM images of TaS_2 reported in air. For examples, Han et al. [6] and Thomson et al. [12] both have investigated the charge-density wave in nearly NC phase of TaS_2 in air. Compared with STM images obtained in UHV condition [13], images taken in air were presented withpoor quality. Among all TMDS, Te-based crystals remained virtually unexplored, especially for MoTe_2, only few STM image taken in vacuum condition was reported [14]. To obtain high quality STM images of TMDS, some studies even need to be carried out under low temperature [14]. A much simpler method to investigate the TMDS surface with atomic resolution is desired for a long time. Cleaving crystals in vacuum chamber is absolutely advantageous to keep the surface fresh, and high quality STM images can be obtained easily. While operations in vacuum, such as cleaving a fresh crystal surface and exchanging the probe tip and sample, are much complicated compared with those in air. Besides, when the size of single crystals is close to micro-meter level, it is impossible to make a probe tip right on the surface center of crystals. Another disadvantage of imaging crystal in vacuum is that new defects and dislocations will be generated for the huge pressure difference between inside and outside of the crystals [15, 16]. In addition to the vacuum protection, liquid condition has also exhibited excellent performance of protecting the fresh crystal surface from being oxidized or adsorbed with impurities. For example, atomic resolution STM images of Au (111) surface can be obtained easily in solution [17, 18], which is not possible in air. If single crystal samples for STM were prepared by cleaving them in liquid directly, even without protection of vacuum, we could easily obtain the high resolution STM images of TMDs. In this work, we have built a STM working in liquid condition with high stability and precision, by cleaving single crystals in our liquid cell to prevent the fresh surface from being oxidized. Using the method above, we have imaged TMDs TiSe_2, MoTe_2, and TaS_2. Single atom defects and triangle defects on TiSe_2 surface were acquired with high resolution. The superstructure of MoTe_2 and CDW structure in NC phase of TaS_2 were also observed at 295 K. Ⅱ. EXPERIMENTS A. Single crystals growth Single crystals of TiSe_2 and TaS_2 were grown by the chemical vapor transport (CVT) method with iodine as a transport agent. The high-purity elements Ti (3.5 N), Se (3.5 N), Ta (3.5 N), and S (3.5 N) were mixed in chemical stoichiometry, and then heated at 800 ℃ for 5 days in an evacuated quartz tube. The obtained TiSe_2/TaS_2 powder and iodine (density: 5 mg/cm^3) were sealed in another longer quartz tube, and then heated for 10 days in a two-zone furnace, where the temperature of source zone and growth zone was fixed at 900 ℃ and 800 ℃, respectively. Single crystal of MoTe_2 was synthesized using the flux method. Well-ground high-purity Mo (3.5 N) and Te (3.5 N) powders were mixed in an appropriate ratio with sodium chloride (NaCl) in an alumina tube. The powder mixture was then sealed in a quartz tube and heated at 1100 ℃ for 12 h, and then cooled down to 900 ℃ at a rate of 0.5 ℃/h. B. STM experiment All experiments were conducted with a home-built STM. The STM consists of three parts: stacked Gecko-Drive piezo motor [19], ultra-compact scanner [20], and sample cell. The stacked Gecko-Drive piezo motor was used for coarse approach by pushing the ultra-compact scanner move forward until the scanning tunneling current was detected. When the coarse approach was done, the piezo motor withdrew and completely detached from the scanner. This separated structure was designed to prevent the disturbance of piezo motor from passing to ultra-compact scanner, thus the junction between tip and sample would be of high stability. The tip was prepared by mechanical shearing an annealed Pt/Ir wire (diameter: 0.25 mm, Pt:Ir=90%:10%, from Alfa Aesar), then coated with poly(methyl styrene) at 180 ℃ to minimize the background current in liquid, coating length of tip was about 5 mm. Sample cell used in our experiment was an O-ring sealed Teflon cell, the depth and capacity of cell were 5 mm and 30 μL respectively. The liquid used in our experiment was Milli-Q ultrapure water (TOC=3 ppb, resistivity=18.2 M\Omega$$\cdot$cm at 25 ℃, purchased from Millipore).

C. Cleaving single crystals

In order to cleave the single crystals with a fresh surface, the samples were processed as seen in FIG. 1. We first coated the oxidized crystal surface with a thin film of Torr Seal (from EI) and attached our T-shaped holder to the Torr Seal-covered crystal surface tightly. The Torr Seal will be dry enough to cleave a fresh crystal surface after 30 min. In order to prevent the fresh surface from being exposed in air, all cleaving process was performed in liquid cell which was filled with ultrapure water. All the STM images were obtained with the constant height imaging mode, and the whole imaging process was finished within 2 h to prevent crystal surface from being oxidized severely.

 FIG. 1 Schematic of cleaving a fresh surface of single crystals in liquid.
Ⅲ. RESULTS AND DISCUSSION A. The design and construction of STM

We have built a high-performance STM working in liquid condition (seen in FIG. 2), the rifting rates in $X$-$Y$ plane and $Z$ direction were as low as 67 pm/min and 55.6 pm/min, respectively. The high precision and stability of our STM have improved imaging quality significantly. Except the probe tip which was exposed to the liquid, the whole STM was fully enclosed to prevent the damages on piezoelectric tube scanner (PTS) caused by the solution evaporation. Using our improved tip coating technology, we could minimize the leakage current to 10 pA. The leakage current is much lower compared with the imaging tunneling current at nA level, thus we could acquire high quality images easily. To check the performance of the STM, we have imaged the highly oriented pyrolytic graphite (HOPG, Grade 2, from SPI) in liquid, high quality atomic resolution image of HOPG was obtained even without "vibration and sound isolations" (FIG. 2(d)). This indicates the high precision and stability of our STM, which was of great importance for imaging the fresh surface of single crystals.

 FIG. 2 STM working in liquid. (a) Photograph of the home-built STM. (b) Stacked Gecko-Drive piezo motor (above) and ultra-compact scanner (below). (c) Exploded view of the sample cell. (d) Constant height STM image of the HOPG taken in liquid without vibration and sound isolations (raw data, bias voltage=+100 mV, scan area=4 nm$\times$4 nm, tunneling current=5 nA).
B. STM investigations of TiSe$_\textbf{2}$ in liquid

Using our home-built STM, we firstly imaged the fresh surface of a cleaved TiSe$_2$ crystal in liquid. To study the (001) surface of TiSe$_2$ crystal consisting of selenium atoms, we cleaved TiSe$_2$ along the van der Waals gap. FIG. 3(a) is the STM image on fresh TiSe$_2$ surface, along with the lattice structure in order, the single atom defect on selenium sites is also displayed. Moreever, defects in the form of triangles rising from surface are clearly seen, consistent with other studies [21, 22]. FIG. 3 (b) and (c) are 3D-images of single atom defect and triangle defect corresponding to the marked areas in FIG. 3(a). FIG. 3(d) is the tunneling current profile along the dashed line shown in FIG. 3(a). As Amzallag and his co-workers suggested [21], defects in FIG. 3 could be related to vacancies or Frenkel defects on titanium sites below the first top layer of selenium atoms, or part of the titanium atoms migrated from their sites in the Se-Ti-Se slab into the interstices of the van der Waals gap. The above defects will affect the local electronic density of the surface as well as its local chemical behavior.

 FIG. 3 STM image of TiS$_2$ single crystals in liquid. (a) STM image of TiS$_2$ fresh surface with scan area 7 nm$\times$7 nm (bias voltage=+100 mV, tunneling current=5 nA). (b) and (c) are 3D-image of single atomic defect and triangular defect marked in (a). (d) Tunneling current profiles along the dashed line in (a).
C. STM investigations of MoTe$_\textbf{2}$ and TaS$_\textbf{2}$ in liquid

Atomic resolution STM image displayed in FIG. 4(a) was obtained on the fresh surface of MoTe$_2$ in 1T phase with a scan area of 1.8 nm$\times$1.8 nm. Surface of 1T-MoTe$_2$ consists of parallel chains of Te atoms and Mo atoms, highly consistent with the observation in films of MoTe$_2$ layers. The bright spots in STM image shown in FIG. 4(a) are due to the Mo atoms at top position, FIG. 4(b) are the top view and side view of 1T-MoTe$_2$ atomic structure respectively.

 FIG. 4 STM image of 1T-MoTe$_2$ single crystals in liquid. (a) STM image of 1T-MoTe$_2$ obtained at $T$=295 K (scan area=1.8 nm$\times$1.8 nm, bias voltage=+100 mV, tunneling current=5 nA). (b) Top view and side view of 1T-MoTe$_2$ atomic structure.

TaS$_2$ is a two-dimensional conductor which supports a rich spectrum of CDW phase, one of the most intriguing CDW structure can be observed at its nearly commensurate phase (NC). It has been established that the NC phase of TaS$_2$ presents a hexagonal domain structure of CDW in which the CDW is commensurate with the atomic lattice. The NC phase exists at the temperature range between 353 and 283 K. Our experiment was carried out at 295 K, thus we could obtain atomic resolution CDW structure of TaS$_2$ crystal in NC phase (seen in FIG. 5(a)), the hexagonal domain structure of the CDW was clearly seen, consistent with the reported studies. FIG. 5(b) is 3D-image of remarked area in FIG. 5(a), and FIG. 5(c) is the two-dimensional Fourier transform power spectra of FIG. 5(a). Besides, we also have obtained CDW images with different shapes, FIG. 6 shows the different CDW structures of TaS$_2$ caused by probe tip.

 FIG. 5 STM image of TaS$_2$ in liquid. (a) STM imagess of TaS$_2$ in NC phase obtained at $T$=295 K (scan area=6 nm$\times$ 6 nm, bias voltage=+100 mV, tunneling current=7 nA). (b) 3D-image of marked area in (a). (c) Two-dimensional Fourier transform power spectra of (a).
 FIG. 6 Different CDW shapes caused by probe tip. STM images of TaS$_2$ obtained at $T$=295 K (bias voltage=+100 mV, tunneling current=7 nA) with scan area (a) 4.5 nm$\times$4.5 nm, (b) 4.5 nm$\times$4.5 nm, (c) 3.8 nm$\times$3.8 nm.

The performance in liquid of our STM illustrates the highly stability and precision. Compared with experiments taken in UHV chamber, the operation in liquid condition is much simple and effective. Operations such as cleaving a fresh crystal surface, exchanging the probe tip and sample, even moving the tip right above a 30 μm$\times$30 μm small sample by means of optical microscope, can be accomplished in 5 min. The simple operation makes the STM experiment more efficient in imaging the lively single crystals. Besides, new defects and dislocations on single crystals will not be generated in liquid as in vacuum.

Ⅳ. CONCLUSION

We have presented a method to study TiSe$_2$, MoTe$_2$ and TaS$_2$ in liquid condition with a home-made STM. The high stability and precision of the STM make it simple and effective to investigate the samples even without any vibration and sound isolations. By cleaving crystal samples in liquid, the fresh surface kept clean in a period time, which was long enough to accomplish the whole STM measurement. Using home-built STM, single atomic defects and triangle defects on TiSe$_2$ crystal surface were acquired with atomic resolution. We also have obtained superstructure of MoTe$_2$ crystal. Besides, the hexagonal charge-density wave domain structure in nearly commensurate (NC) phase of TaS$_2$ was also observed at 295K. The method of cleaving and imaging the lively samples in liquid with high-stable STM hasshown the great potential in investigating the atomic structure and defects on lively sample surface.

Ⅴ. ACKNOWLEDGMENTS

We sincerely thank Prof. Y. P. Sun at High Magnetic Field Laboratory (CAS) for providing the TMDS samples. This work was supported by the National Key R & D Program of China (No.2017YFA0402903 and No.2016YFA0401003), the National Natural Science Foundation of China (No.11804345, No.U1632160, No.51627901, No.21505139, No.11704384), Chinese Academy of Sciences Scientifc Research Equipment (Grant YZ201628), the Anhui Provincial Natural Science Foundation (No.1808085MB51, No.1608085MB36), the Innovative Program of Development Foundation of Hefei Center for Physical Science and Technology (No.2018CXFX001), the Dean fund of Hefei Institutes of Physical Science of CAS (Grant YZJJ201620).

Reference
 [1] S. Z. Butler, S. M. Hollen, L. Y. Cao, Y. Cui, J. A. Gupta, H. R. Gutierrez, T. F. Heinz, S. S. Hong, J. X. Huang, A. F. Ismach, E. Johnston-Halperin, M. Kuno, V. V. Plashnitsa, R. D. Robinson, R. S. Ruoff, S. Salahuddin, J. Shan, L. Shi, M. G. Spencer, M. Terrones, W. Windl, and J. E. Goldberger, ACS Nano 7 , 2898 (2013). DOI:10.1021/nn400280c [2] M. Lavarone, R. D. Capua, X. Zhang, M. Golalikhani, S. A. Moore, and G. Karapetrov, Phys. Rev. B 85 , 1903 (2012). [3] J. C. E. Rasch, T. Stemmler, B. Müller, L. Dudy, and R. Manzke, Phys. Rev. Lett. 101 , 237602 (2008). DOI:10.1103/PhysRevLett.101.237602 [4] D. Greenaway, and R. Nitsche, J. Phys. Chem. Solids 26 , 1445 (1965). DOI:10.1016/0022-3697(65)90043-0 [5] R. E. Thomson, B. Burk, A. Zettl, and J. Clarke, Phys. Rev. B 49 , 16899 (1994). DOI:10.1103/PhysRevB.49.16899 [6] R. E. Thomson, U. Walter, E. Ganz, J. Clarke, A. Zettl, P. Rauch, and F. J. Disalvo, Phys. Rev. B 38 , 10734 (1988). DOI:10.1103/PhysRevB.38.10734 [7] A. Lepetit, J. De Physique 26 , 175 (1965). DOI:10.1051/jphys:01965002604017500 [8] M. Morsli, A. Bonnet, V. Jousseaume, L. Cattin, A. Conan, and M. Zoaeter, J. Mater. Sci. 32 , 2445 (1997). DOI:10.1023/A:1018569510512 [9] J. T. Wang, W. F. Ge, Y. B. Hou, and Q. Y. Lu, Carbon 84 , 74 (2015). DOI:10.1016/j.carbon.2014.11.046 [10] H. S. Wong, and C. Durkan, Nanotechnology 23 , 185703 (2012). DOI:10.1088/0957-4484/23/18/185703 [11] B. Hildebrand, C. Didiot, A. M. Novello, G. Monney, A. Scarfato, A. Ubaldini, H. Berger, D. R. Bowler, C. Renner, and P. Aebi, Phys. Rev. Lett. 112 , 1495 (2014). [12] W. Han, R. A. Pappas, E. R. Hunt, and F. R. Frindt, Phys. Rev. B 48 , 8466 (1993). DOI:10.1103/PhysRevB.48.8466 [13] P. Schmidt, J. Kröger, B. M. Murphy, and R. Berndt, New J. Phys. 10 , 773 (2008). [14] I. G. Lezama, A. Ubaldini, M. Longobardi, E. Giannini, C. Renner, A. B. Kuzmenko, and A. F. Morpurgo, 2D Materials 1 , 021002 (2014). DOI:10.1088/2053-1583/1/2/021002 [15] J. Jung, Philos. Mag. A 50 , 257 (1985). DOI:10.1080/01418618408244226 [16] I. M. Shmyt'ko, E. A. Kudrenko, V. V. Sinitsyn, B. S. Red'kin, and E. G. Ponyatovsky, Jetp Lett. 82 , 409 (2005). DOI:10.1134/1.2142866 [17] Z. G. Xia, J. H. Wang, Y. B. Hou, and Q. Y. Lu, Rev. Sci. Instrum. 85 , 096103 (2014). DOI:10.1063/1.4894470 [18] Z. G. Xia, J. H. Wang, Y. B. Hou, and Q. Y. Lu, Rev. Sci. Instrum. 85 , 125103 (2014). DOI:10.1063/1.4902975 [19] Q. Wang, Y. B. Hou, and Q. Y. Lu, Rev. Sci. Instrum. 84 , 056106 (2013). DOI:10.1063/1.4805058 [20] Q. Wang, Y. B. Hou, J. T. Wang, and Q. Y. Lu, Rev. Sci. Instrum. 84 , 113703 (2013). DOI:10.1063/1.4829716 [21] E. Amzallag, I. Baraille, H. Martinez, M. Ré rat, and D. Gonbeau, J. Chem. Phys. 128 , 014708 (2008). DOI:10.1063/1.2814243 [22] G. P. Van Bakel, and J. T. De Hosson, Phys. Rev. B 466 , 2001 (1992).

a. 中国科学技术大学合肥微尺度物质科学国家研究中心，合肥 230026;
b. 中国科学院强磁场科学中心极端条件凝聚态物理安徽省重点实验室，合肥 230031;
c. 中国科学院合肥大科学中心，合肥 230031;
d. 南京大学人工微结构科学与技术协同创新中心, 南京 210093