Ohmic Contact Fabrication Using a Focused-ion Beam Technique and Electrical Characterization for Layer Semiconductor Nanostructures

Summary

We describe the approaches for the device fabrication and electrical characterization of molybdenum diselenide (MoSe₂) layer semiconductor nanostructures with different thicknesses. In addition, the fabrication of ohmic contacts for MoSe₂-layer nanocrystals by the focused-ion beam deposition method using platinum (Pt) as a contact metal is described.

Abstract

Layer semiconductors with easily processed two-dimensional (2D) structures exhibit indirect-to-direct bandgap transitions and superior transistor performance, which suggest a new direction for the development of next-generation ultrathin and flexible photonic and electronic devices. Enhanced luminescence quantum efficiency has been widely observed in these atomically thin 2D crystals. However, dimension effects beyond quantum confinement thicknesses or even at the micrometer scale are not expected and have rarely been observed. In this study, molybdenum diselenide (MoSe₂) layer crystals with a thickness range of 6-2,700 nm were fabricated as two- or four-terminal devices. Ohmic contact formation was successfully achieved by the focused-ion beam (FIB) deposition method using platinum (Pt) as a contact metal. Layer crystals with various thicknesses were prepared through simple mechanical exfoliation by using dicing tape. Current-voltage curve measurements were performed to determine the conductivity value of the layer nanocrystals. In addition, high-resolution transmission electron microscopy, selected-area electron diffractometry, and energy-dispersive X-ray spectroscopy were used to characterize the interface of the metal–semiconductor contact of the FIB-fabricated MoSe₂ devices. After applying the approaches, the substantial thickness-dependent electrical conductivity in a wide thickness range for the MoSe₂-layer semiconductor was observed. The conductivity increased by over two orders of magnitude from 4.6 to 1,500 Ω⁻¹ cm⁻¹, with a decrease in the thickness from 2,700 to 6 nm. In addition, the temperature-dependent conductivity indicated that the thin MoSe₂ multilayers exhibited considerably weak semiconducting behavior with activation energies of 3.5-8.5 meV, which are considerably smaller than those (36-38 meV) of the bulk. Probable surface-dominant transport properties and the presence of a high surface electron concentration in MoSe₂ are proposed. Similar results can be obtained for other layer semiconductor materials such as MoS₂ and WS₂.

Introduction

Transition metal dichalcogenides (TMDs), such as MoS₂, MoSe₂, WS₂, and WSe₂, have an interesting two-dimensional (2D) layer structure and semiconducting properties^1-3. Scientists have recently discovered that the monolayer structure of MoS₂ shows a substantially enhanced light-emitting efficiency because of the quantum confinement effect. The finding of the new direct-bandgap semiconductor material has attracted substantial attention^4-7. In addition, the easily stripped layer structure of TMDs is an excellent platform for studying the fundamental properties of 2D materials. Unlike metallic graphene without the bandgap, TMDs have inherent semiconducting characteristics and have a bandgap in the range of 1-2 eV^1,3,8. The 2D structures of the ternary compounds of TMDs⁹ and the possibility of the integration of these compounds with graphene provide an unprecedented opportunity to develop ultrathin and flexible electronic devices.

Unlike graphene, the room temperature electron mobility values of 2D TMDs are at a moderate level (1-200 cm²V⁻¹sec⁻¹ for MoS₂^10-17; approximately 50 cm²V⁻¹sec⁻¹ for MoSe₂¹⁸). The optimal mobility values of graphene have been reported to be higher than 10,000 cm²V⁻¹sec⁻¹.^19-21 Nevertheless, semiconducting TMD monolayers exhibit excellent device performance. For instance, the MoS₂ and MoSe₂ monolayers or multilayer field-effect transistors exhibit extremely high on/off ratios, up to 10⁶-10^{9 10,12,17,18,22}. Therefore, it is crucial to understand the fundamental electrical properties of the 2D TMDs and their bulk materials.

However, studies of the electrical properties of the layer materials have been partially hampered because of the difficulty in forming good ohmic contact on the layer crystals. Three approaches, shadow mask deposition (SMD)²³, electron beam lithography (EBL)^24,25, and focused-ion beam (FIB) deposition,^26,27 have been used to form electrical contacts on nanomaterials. Because SMD typically involves the use of a copper grid as the mask, the spacing between two contact electrodes is mostly larger than 10 μm. Unlike EBL and FIB deposition, metal deposition of electrode arrays on a substrate is performed without targeting or selecting nanomaterials of interest in the SMD method. This approach cannot guarantee that the metal patterns are correctly deposited on individual nanomaterials as the electrodes. The result of the SMD method has an element of chance. The EBL and FIB deposition methods are used in the scanning electron microscope (SEM) system; nanomaterials can be directly observed and selected for electrode deposition. In addition, EBL can be used to easily fabricate metal electrodes with a line width and a contact electrode spacing smaller than 100 nm. However, the residual resist on the nanomaterial surface left during lithography inevitably results in the formation of an insulating layer between the metal electrode and the nanomaterial. Thus, EBL leads to high contact resistance.

The main advantage of electrode fabrication through FIB deposition is that it leads to low contact resistance. Because metal deposition is performed by the decomposition of an organometallic precursor by using an ion beam at the defined area, metal deposition and ion bombardment occur simultaneously. This could destroy the metal–semiconductor interface and prevent the formation of Schottky contact. Ion bombardment can also eliminate surface contaminants such as hydrocarbons and native oxides, which decreases contact resistance. Ohmic contact fabrication through FIB deposition has been demonstrated for different nanomaterials^27-29. In addition, the entire fabrication procedure in the FIB deposition approach is simpler than that in EBL.

As layer semiconductors typically show highly anisotropic electrical conduction, the conductivity in the layer-to-layer direction is several orders of magnitude lower than that in the in-plane direction^30,31. This characteristic increases the difficulty of fabricating ohmic contacts and determining electrical conductivity. Therefore, in this study, FIB deposition was used for studying the electrical properties of layer semiconductor nanostructures.

Protocol

1. Structural Characterization of MoSe₂ Layer Crystals (See Step 1 in Figure 1)

1. XRD Measurement Procedure
   1. Mount a MoSe₂ layer crystal (with the size range of 5 x 5 x 0.1-10 x 10 x 0.5 mm³) or crystal powder (which was mixed with quartz powder and binder and was smeared on the slide glass) on the holder.
   2. Press the holder by a slide glass to ensure layer crystal surface parallel to the holder surface.
   3. Load the sample holder into the diffractometer.
   4. Close the doors of diffractometer.
   5. Calibrate beam line accordingly to manufacturer’s instructions.
   6. Input measurement parameters such as the 2 scan range (10-80°), the increment (0.004°), and the dwell time (0.1 sec).
   7. Start the DIFFRAC.Measurement Center program on the computer attached to the diffractometer and then save data and name data file according to manufacturer’s protocol.
   8. Analyze the XRD pattern by identifying the positions of the diffraction peaks using the software and then compare with the standard data from JCPDS card database to confirm the single out-of-plane orientation and single-crystalline quality of the MoSe₂ layer crystals^32,33.
2. Micro-Raman Measurement Procedure
   1. Perform the Raman equipment calibration using a silicon wafer as the standard sample. The measurement of the silicon wafer is the same as the procedure described below for the interested MoSe₂ layer crystal.
   2. Mount a MoSe₂ layer crystal on the slide glass.
   3. Load the slide glass on the holder of optical microscope and focus the sample surface with a white light source.
   4. Switch the light source from a white light to a laser beam (wavelength at 514 nm).
   5. Input measurement parameters such as the wavenumber scan range (150-500 cm^-1), the integration time (10 sec), and the number of scan times (10-30 times).
   6. Start the program on computer attached to the Raman spectrometer and then save data and name data file according to manufacturer’s protocol.
   7. Analyze the Raman spectrum by identifying their peak widths and positions using the software and then compare with standard data from references to confirm the crystalline structure type and quality of the MoSe₂ layer crystals^34,35.

2. Fabrication of MoSe₂ Layer Nanocrystal Devices

1. Mechanical Exfoliation of Layer Crystals
   1. Clean tweezers with acetone and alcohol.
   2. Pick the MoSe₂ layer crystals (4 to 8 pieces) with a shiny surface (i.e. mirror-like crystal face) and an area size larger than 0.5 x 0.5 mm² with the tweezers and put them on the dicing tape with an area size of 20 x 60 mm².
   3. Fold the tape in half to exfoliate the layer crystal and repeat the action approximately twenty times. Usually layer crystals can be stripped into many micrometer-sized crystals in width (see Step 2 in Figure 1).
   4. Load the dicing tape with the layer nanocrystal powder into the SEM chamber to observe the sizes and morphologies of these stripped MoSe₂ layer microcrystal. If the width distributions of the layer nanocrystal are at 1-20 μm, the nanocrystal powder can meet the criteria for the device fabrication.
2. Dispersion of the Layer Nanocrystals on the Device Template
   1. Place the dicing tape with the layer nanocrystal powder upside-down on the device template. The template is SiO₂ (300 nm)-coated silicon substrate with sixteen pre-patterned Ti (30 nm)/Au (90 nm) electrodes on the SiO₂ surface (see Step 4 in Figure 1). The area size of the template is 5 x 5 mm².
   2. Tap the dicing tape lightly to make some nanocrystals (roughly 10 to 100 pieces) fall on the template.
   3. Check the number density and dispersion condition of the nanocrystal on the template by optical microscope or sometimes by SEM if the dispersed nanocrystals can be not observed by optical microscope. Usually 2 to 5 pieces of nanocrystals (area size larger than 2 x 2 μm²) dispersed on the center square (with area of 80 x 80 μm²) of the template without overlapping to each other are the better condition for the next FIB processing.
3. Electrode Fabrication by FIB
   1. Mount templates on the FIB holder using conducting copper foil tape. Typically, the area of conducting tape of 3 x 2.4 cm² was required for mounting 6-8 templates.
   2. Load the holder into the FIB chamber.
   3. Evacuate the chamber to the vacuum degree down to 10^-5 mbar by clicking the button “Pump”.
   4. Set the electron beam current (41 pA) and acceleration voltage (10 kV) for the SEM mode.
   5. Set the ion beam current (0.1 nA) and acceleration voltage (30 kV) for the FIB mode.
   6. Warm up the ion beam system and gas-injection-system (GIS) by clicking the button “beam on” and the button “Cold” in the “Gas Injection” block, respectively.
   7. Turn on the electron-beam by clicking the button “Beam On” and focus the image at a low magnification of 100X.
   8. Set the z-axial working distance (WD) at 10 mm for SEM mode.
   9. Set the magnification at 5,000X and focus.
   10. Set the tilt angle of the holder to 52 degrees by clicking the button “Navigation” and input the tilt angle “52”.
   11. Select a MoSe₂ layer nanocrystal with a certain thickness (ranging from 5 to 3,000 nm) and a rectangular and square shape for the electrode fabrication.
   12. Take the SEM images at different magnifications (from 1,000X to 10,000X) of the targeted pristine material before electrode fabrication by clicking the button “Snapshot”.
   13. Switch to FIB mode and take an FIB image by the snapshot mode to reduce the exposure time of the targeted material under ion beam bombardment.
   14. Define the electrode deposition area, select the “Pt deposition” mode, and input the thickness (0.2-1.0 μm) value of the deposited Pt electrode.
   15. Introduce the capillary of GIS into the chamber by clicking the box “Pt dep” in the “Gas Injection” block.
   16. Take an image by the snapshot mode again and modify the position of the electrodes if the originally defined pattern shifts slightly.
   17. Turn on the FIB deposition by clicking the button “Start Patterning”.
   18. After deposition, draw the capillary of GIS back by unclicking the box “Pt dep” in the “Gas Injection” block.
   19. Switch to SEM mode and check the result of the deposited Pt electrodes on the layer nanocrystal.
   20. Take the SEM images at different magnifications of the completed devices with two or four electrodes (see Step 3 in Figure 1).
   21. Set the tilt angle of the holder return to 0 degrees by clicking the button “Navigation” and input the tilt angle “0”.
   22. Take the top-viewed SEM images at different magnifications for the estimations of the material width and electrode inter-distance by clicking the button “Snapshot”.
   23. Turn off the electron-beam and ion beam systems and cool down GIS system by clicking the button “Beam Off” and the button “Warm” in the “Gas Injection” block, respectively.
   24. Vent the chamber by introducing nitrogen gas by clicking the buttons “Vent” and then take the holder out of the chamber. It typically takes 5 to 10 min to finish the venting process.
   25. Close the chamber door and evacuate the chamber.

3. Characterizations of the MoSe₂ Layer Nanocrystal Devices

1. Thickness Measurement of the Layer Nanocrystals by AFM
   1. Install the AFM cantilever to the probe holder.
   2. Turn on AFM program and select “ScanAsyst” mode.
   3. Load the probe holder and connect it with the laser diode head of AFM station.
   4. Perform the calibration to align the incident laser beam position and the cantilever according to manufacturer’s protocol.
   5. Mount the sample (the template chip with FIB-fabricated layer nanocrystal devices) on the sample holder by Cu foil tape.
   6. Load the sample holder to the AFM station.
   7. Move the sample holder to the position approximately underneath the laser beam or AFM cantilever.
   8. Lower down AFM cantilever to the focus position by focusing the optical microscope image of the layer nanocrystal.
   9. Input scan parameters such as the scan area (6 x 6-30 x 30 μm²), the frequency (0.5-1.5 Hz), and the resolution (256-512 lines).
   10. Start the program and save data according to the manufacturer’s protocol.
   11. Raise the AFM cantilever and take the sample holder out.
   12. Load the second sample and repeat the measurement procedure described above if need.
   13. Estimate the thickness of layer nanocrystals by analyzing the AFM image and height profile using the software “NanoScope Analysis”. Select a lateral height profile from the AFM image and determine the average thickness value by the flatten area of the profile. (See Figure 2d and 2e)
2. Current versus voltage (I-V) measurement of the layer nanocrystals
   1. Mount the sample (the template chip with FIB-fabricated layer nanocrystal devices) on the mica substrate by Cu foil tape.
   2. Bond the enameled wires or Cu wires on the electrodes of the chip by Ag paste. (See Step 4 in Figure 1.)
   3. Load the completed sample in the probe station chamber and fix it on the sample holder by Cu foil tape. The cryogenic probe station was located in the dark environment. (See Step 5 in Figure 1.)
   4. Solder the electrical wires of the sample and the metal electrodes of the probes one by one.
   5. Cap the chamber top and evacuate the chamber down to 10^-4 mbar. Cool down sample to the 77 K by introducing liquid nitrogen into the probe station. Set the temperature range (usually from 80 to 320 K), interval, and dwell time for the temperature control. (Necessary only for temperature-dependent measurement).
   6. Set the applied voltage sweeping range (typically from -1 to 1 V), voltage interval (0.01 V), and the limited maximal current (10 or 100 μA) in an ultrahigh-impedance multifunctional electrometer for the two-terminal I-V measurement. For four-terminal measurement, set the applied current sweeping range (typically from -100 to 100 μA) and current interval (1 μA).
   7. Start the program and save the I-V data at room temperature or at different temperatures.
   8. Open the chamber cap if necessary and take the sample out of the chamber.
   9. Load the second sample if need and repeat the procedure described above.
   10. Analyze the I-V curve by plotting the measured current versus applied voltage data using the software. Fit the I-V curve by selecting the Linear Fitting function. Check the linearity of the I-V curve and obtain the slope value (i.e. conductance value). (See Step 6 in Figure 1.)
   11. Repeat the Step 3.2.10 for the I-V curves measured at different temperatures if need.
   12. Calculate the conductivity (σ) value according to the equation σ= G(t/tw) by adopting the parameters obtained by I-V, SEM, and AFM measurements including conductance (G), thickness (t), width (w) and length (l) of the layer nanocrystal.
   13. Plot the curves of the conductance and conductivity values versus thickness of layer nanocrystals.

Results

The determined values of the electrical conductance (G) and conductivity (σ) of layer nanomaterials with different thicknesses are highly dependent on the quality of the electrical contacts. The ohmic contacts of the FIB-deposition-fabricated two-terminal MoSe₂ devices are characterized by measuring the current–voltage (I–V) curve. The room temperature I–V curves for the two-terminal MoSe₂ nanoflake devices with different t...

Discussion

The accurate determination of the σ value and its dimension dependence in the layer nanocrystals is highly dependent on the quality of the electrical contacts. The FIB deposition method used for metal electrode deposition played a crucial role throughout the study. According to electrical, structural, and composition analyses, the fabrication of stable and highly reproducible ohmic contacts, using FIB deposition method, in the MoSe₂ or MoS₂ devices was facilitated by the formation of t...

Materials

Name	Company	Catalog Number	Comments
HRTEM&SEAD	FEI (http://www.fei.com/products/tem/tecnai-g2/?ind=MS)	Tecnai™ G2 F-20
SEM&EDS	HITACHI (http://www.hitachi-hitec.com/global/em/sem/sem_index.html)	S-3000H
FIB	FEI (http://www.fei.com/products/dualbeam/versa-3d/)	Quanta 3D FEG
AFM	BRUKER (http://www.bruker.com/products/surface-analysis/atomic-force-microscopy/dimension-icon/overview.html)	Dimension Icon
XRD	Bruker (https://www.bruker.com/products/x-ray-diffraction-and-elemental-analysis/x-ray-diffraction/d2-phaser/learn-more.html)	D2 PHASER X-ray Diffractometer
Raman	Renishaw (http://www.renishaw.com/en/renishaw-enhancing-efficiency-in-manufacturing-and-healthcare--1030)	inVia Raman microscope system
Keithley-4200	keithley (http://www.keithley.com.tw/products/dcac/currentvoltage/4200scs)	4200scs
ultralow current leakage cryogenic probe station	Lakeshore Cryotronics (http://www.lakeshore.com/)	TTP4
copper foil tape	3M (http://solutions.3m.com/wps/portal/3M/en_US/Electronics_NA/Electronics/Products/Product_Catalog/~/3M-Copper-Foil-Shielding-Tape-1182?N=4294300025+5153906&&Nr=AND%28hrcy_id%3A8CQ27CX0WMgs_F2LMWMM6M6_N2RL3FHWVK_GPD0K8BC31gv%29&rt=d)	1182
Ag paste	Well-Being (http://www.gredmann.com/about.htm)	MS-5000
Cu wire	Guv Team (http://www.guvteam.com)	ICUD0D01N
dicing tape	Nexteck (http://www.nexteck-corp.com/tw/product-tape.html)	contact vender
mica	Centenary Electronic (http://100y.diytrade.com/sdp/307600/4/pl-1175840/0.html)	T0-200
enamel wire	Light-Tech Electronics (http://www.ltc.com.tw/product_info.php/products_id/57631)	S.W.G #38