Essay: Germanium and MoS2 Device Characterization

The success of Silicon based technology is due to SiO2/Si interface. In this, the interface defect density is about 1012 cm-2 but it can be reduced by two orders after annealing the device in N2/H2 atmosphere at temperature nearly 400-4500C. In case of Germanium, the oxide/Ge interface density is about 1013cm-2 ranges which harm the electrical characteristics of MOS devices. Also its oxide is thermodynamically unstable in the range 400-450 0C. For this reason it is not used generally. Recently, this problem has been overcome by using high-” dielectrics. In this, we study about the Germanium MOSCAP and Germanium MOSFET and also about its parameters and characteristics. In this, we can extract the MOSCAP capacitance-voltage characteristics and compare the CV of implanted and without implanted samples and also see the effect of annealing on CV characteristics of MOSCAP. For MOSFET, we calculate its mobility using Split CV technique and series source-drain resistance with varying channel length.
The two-dimensional material MoS2 is also now mostly used because of their low dimensionality, excellent physical, electrical and optical properties. These devices have wide applications from conventional electronic and optoelectronic devices to new fields like spintronics, valleytronics and straintronics. These reports explore the physical and electrical properties of materials and also the MoS2 FET and its characteristics.
CHAPTER 1
INTRODUCTION
1.1 GENERAL OVERVIEW
The downscaling trend of Complementary Metal Oxide Semiconductor (CMOS) field effect transistor will continue as its decreases cost per function in a chip, decreases power consumption and increases performance. To reduce the leakage through gate oxide high-” dielectrics are used. From past few years the native oxide on Ge are unstable which is the biggest problem for the integration of CMOS devices on large scale. With high-” dielectrics the Ge/high-” interface still remains a problem if GeO2 is not controlled properly [4].
During early stages, devices are made by depositing the high-” dielectric HfO2 directly on Ge which shows the hysteresis which is mainly due to formation or growth of unstable interfacial layer of GeO2 during or after deposition of HfO2. Diffusion of Ge into HfO2 is also a problem. To achieve excellent electrical performance of Ge devices with high-” dielectrics more research is carried out on the gate stacks. When GeOxNy is deposited prior to HfO2 it show high mobility but the interface problem still remains the same. We also use the different gate stacks of Ge oxynitride by using thermal or anodic nitridation and also by using GeON as low temperature for Ge MOSFET but they are not scalable. In other engineering treatment the surface-nitridation is done before HfO2 growth by exposing the surface of Ge to an atomic N beam from a remote RF source. The significant dispersion in inversion region is observed as the frequency reduces which shows that there is slow interface states exists. Surface treatment of Ge prior to oxide deposition are seems to improve the MOS device quality but it is not for integration. Recently, it has been observed that when the GeO2 is grown on p-type Ge surface by thermal oxidation an improved GeO2/Ge interface is observed with defect density nearly equal to 1011 cm-2eV-1 without any annealing or interfacial passivation. The permanent GeO2 layer is required to integrate Ge into standard CMOS technology. After research from many years it has found that HfO2/Al2O3/GeO2/Ge is the best dielectric stack for MOS devices. In this report, chapter 2 is Germanium MOSCAP in which we discuss about the fundamentals of MOSCAP, its working, parameters and characteristics and also see the effect of implantation in MOSCAP which we done to reduce the leakage current. The chapter 3 is Germanium MOSFET, in this we discuss about the MOSFET fundamentals, its working and transfer characteristics and also find the mobility by two ways and compare it to see the difference and also calculate the series source-drain resistance with varying channel length.
A whole new class of materials is obtained if the 2D material is obtained from layered material. These materials have their peculiar structure, extraordinary electrical, physical and optical properties. 2D materials are not present in nature as 2D sheets. Graphene is the pioneer 2D material that burst the research interest in 2D material in the whole worldwide. This new class of 2D material can be represent as Transition Metal Dichalcogenides (TMDC) composition- MX2 where M is transition metal and X is chalcogen. Mo and W are the most widely used transition metals and S, Se and Te are the commonly used chalcogen. TMDCs can be used as semiconductors, metals/semimetals or also as superconductors. There is d-orbital’s which are unsaturated and results in distinct properties of TMDCs and the engineering of filling these orbital’s allow us to design material properties like band gap, conductivity. The existence of inorganic compound MoS2 which is extracted from molibdenite used in lubrication and petroleum refining. This is the material which can compete with today’s novel devices in electronics. TMDCs have strong molecular intralayer bonds and weak interlayer that gives rise to its layered structure and also to consequent anisotropic properties. In this report, chapter 4 concentrates on the MoS2 FET in which we discuss its transfer characteristics [24].
1.2 GERMANIUM
In the late 1940s, two elements silicon and germanium are discovered which acts as a semiconductors i.e. to conduct electricity more efficiently than insulators but less efficient than conductors [1]. Both silicon and germanium are used as semiconductors in electronics. Silicon is the second abundant element on the earth’s crust which found in almost all rocks, all natural water, plants and animal skeleton, as well as body tissues and fluids. Germanium, on other hand is the rare material, which is only 0.0005% of earth’s crust making its recovery difficult. Germanium is not found in natural concentrations or deposits, but rather it is evenly distributed in trace quantities in many rocks. Similar to Si, Ge is also found in various uncommon materials and sulphide ores, especially in silver, tin, lead, zinc and antimony. The basic properties of germanium are shown in Table 1.1.
Lattice constant (”) at 300 0K
5.64613
Crystal Structure
Diamond
Bandgap (eV)
0.66
Permittivity
16
Electron Mobility(cm2/Vs)
3900
Hole Mobility ( cm2/Vs)
1900
Table 1.1: Basic Properties of Germanium
1.2.1 Occurrence of Elemental Germanium
In 1886, Clemens Winkler analyzed the argyrodite, 4AgS.GeS2, a sulphide mineral and found a previously unknown constituent; he isolated it and named ‘Germanium’ after his native country of Germany. This mineral is comparatively rare and not become an important source. Another sulphide mineral germanite, 7CuS.FeS.GeS2, which is found in South-West Africa, contains about 6% of germanium, but its value as a source of commercial quantities of germanium is still in doubt. Most of the Ge produced in United States comes from smelting of Ge rich zinc ores mines in Tri-state district of Missouri, Kansas, and Oklahoma. The zinc sulphide in these ores contains 0.01 to 0.1 % of germanium and Ge is obtained as a by-product of zinc smelting. Recently, commercially available germanium sulphide, renierite which yields 6.4 to 7.8 % of germanium, has been discovered in Prince Leopold copper mine at Kipushi in the Belgian Congo. Coal Ashe and flue dusts are another potential for germanium. Further investigation by the chemists confirms its place between silicon and tin in group IVA in the periodic table. Toward the World War II, new processing methods are made available to produce large quantities of pure germanium which opens the door to use it in semiconductor industry. Highly pure germanium ingots are remelted and doped with specific impurities to produce desired electronics characteristics. About 1945’s Germanium properties are widely known as semiconductor. The first transistor was made in 1947 and after a year the germanium triodes come into the market. After this, diodes, rectifiers, vaccum tubes had been made with germanium [2].
1.2.2 Physical Properties of Germanium
Elementary germanium is silver- white metal having much the outward appearance of bright solder when fused. Germanium is very hard so it has a tendency to fracture when drilled or cut. When the germanium is fractured, the break is shiny and hard similar to metallic silicon. Like bismuth and gallium it expands on freezing [2].
Atomic Number
34
Specific heat (cal/g)
0.074
Atomic Weight
72.60
Hardness (Moh’s scale)
6.0 ‘ 6.5
Density (g/ml, 20 0C)
5.36
Electrical resistivity (‘/cm)
0.089
Melting Point (0C)
958.5
Crystal form Face- centred cubic
Boiling Point (0C)
2700
Atomic Volume (ml/mole)
13.2
Table 1.2: Physical Properties of Germanium
Germanium metal is not appreciably affected by water, 50 per cent sodium hydroxide, concentrated hydrochloric acid, dilute hydrochloric acid or sulphuric acid. It is tarnished by 10 percent sodium hydroxide and nitric acid. There is slight corrosion is observed with sulphuric acid and hydrofluoric acid. It can be dissolved by 3 percent hydrogen- peroxide. The metal is volatile at low pressure at 760 0C. It will absorb 0.186ml of hydrogen per gram of metal when melted and cooled in that gas. It is brittle rather than ductile, with the arrangement of atoms in its crystalline structure similar to diamond lattice of carbon. Before its use in semiconductor industry it had commercial importance. It had been used for dental work in an alloy, in glass, for infrared optical devices such as lenses, prisms and windows.
1.2.3 Atomic Structure of Germanium
Germanium is a substance which consists of atoms all having same number of protons. The atomic number of germanium is 32 so it has 32 protons, 41 neutrons and four valence electrons. So, it has electronic configuration 1s2, 2s2 2p6, 3s23p63d10, 4s24p2.
(a) (b)
Fig.1.1: (a) Germanium, (b) Germanium atomic structure
1.2.4 Energy Band Structure of Germanium
Germanium is an indirect band gap material with Eg = 0.66eV which lies in infrared region. Its band gap structure is shown in Fig1.2. The conduction minima is in (111) direction. The germanium band gap is less as compared to silicon.
Fig 1.2: Energy Band Diagram of Germanium [3]
1.2.5 Electrical Properties of Elemental Germanium
The electrical property of germanium that receives most of the attention at present is its semiconductor property. Germanium diodes and triodes are already replacing the conventional vaccum tubes in every installation [2].
Germanium can be easily deposited as a film on glass, quartz or on dense ceramic materials by the decomposition of monogermane at temperature above 370 0C. Practically, this can be used in the formation of resistors. Resistances measured over 2.5 cm length of germanium deposited inside 7mm tubing range from about 100 ohms to several megohms, which depends on the condition of the deposition. These resistors have low temperature coefficients which vary from 0.0001 to 0.0003 ohms/0C. If we deposited the germanium film over silver, resistors with low resistance values but temperature coefficients of less than 0.0001 ohms/0C are obtained. Germanium powder compacts show microphone property properties which is that the resistance varies considerably with external pressure.
1.2.6 Germanium Surface Passivation
The success of Si based technology is due to good quality of SiO2/Si interface. The density of interface defect is ~1012cm-2 which can be decreased two orders by annealing the device in N2/H2 atmosphere. However, in case of Germanium it is about 1013 cm-2 and also it is thermodynamically unstable at 400-450 0C. Such large defect density at the interface will affect the MOS devices performance. Now, this problem had been overcome by using the high-” dielectrics. Recent progress has been achieved regarding Ge surface passivation by Si capping layer, thermally grown GeO2 or GeON interfacial layer or PH3 or H2S surface treatment. This surface treatment combined with high-” dielectric gives the promising device characteristics [4]. There are following types of passivation:
1.2.6.1 Epitaxial Si layer Passivation
In this approach, the ultrathin Si epitaxial layer is deposited which is followed by chemical or thermal oxidation. Ge surface is first etched with HF solution to reduce native oxide GeOx layer. The wafer is then transferred to epitaxial reactor where it is baked with H2 atmosphere temperature of 600-6500C for obtaining the clean Ge surface. After this, ultrathin Si layer is deposited on Ge either using SiH4 at 500 0C or Si3H8 at 350 0C. This interlayer is then oxidized in ozone-based solution to form an ultra thin SiO2 interlayer. By using this interlayer the capacitance-voltage characteristics are so much improved. Now the interface defect density reduces to 2 x1011 cm-2[4].
1.2.6.2 Passivation of Ge with GeO2
Although GeO2 converts to GeO at 430 0C yet it is used for Ge surface passivation. If GeO2 is capped with metal or dielectric it provides robustness to GeO deabsorption. The thermal oxidation is done either with dry O2, ozone, atomic oxygen. However, here we only discuss about dry O2 in temperature between 350 to 450 0C which is followed by Al2O3, ZrO2 and HfO2 deposition using ALD. The interface defect density reduces to 1011 cm-2 near midgap [4].
1.2.6.3 Rare Earth oxides on Ge
High-” dielectric HfO2 or ZrO2 are used in Ge based MOSFETs. Among this, rare earth based oxides such as La2O3 and CeO2 receives attention because they are directly deposited on the Ge which provides efficient passivation to Ge surface. The deposition of these oxides promotes the formation of stable germinate interlayer which is responsible for Ge surface passivation. MOS structures with Pt/La2O3/Ge gate stacks show good CV characteristics with minimum frequency dispersion in accumulation and depletion. These earth oxide/Ge stacks also has been successfully integrated in MOSFETs with good electrical characteristics [4].
1.2.7 Source-Drain Contacts on Germanium
As the dimension of MOSFET reduces to nanoscale the contact to source and drain should be ohmic to make full advantage of device performance [6]. Also germanium channels are used to increase the mobility. We need that the specific contact resistance to both source and drain should be small and also they are able to supply large current so that the voltage drop through it is small. In addition, for short channel devices with high ON currents, the sheet resistance of metal Rsh in contact with heavily doped source and drain should be small to minimize series resistance. In this, we discuss about different types of contacts:
1.2.7.1 Metal/Semiconductor Contacts
The resistance of metal is related to energy barrier or Schottky barrier that exists at M-S interface. Barrier height of electrons depends on metal work function and semiconductor affinity. So, we choose metal in such a way that Fermi levels of both sides are at same level and no barrier exists so that carrier flow across interface and lead to low resistivity ohmic contact. Sometimes heavy doping of semiconductor is used to provide tunnelling through barrier and lower ohmic contact resistance but Schottky barrier height reduction is also desirable [6]. The pinning factor is given by:
S = ( ”_be)/(”_m ) (1.1)
In weak pinning, S=1 and in strong pinning S=0. Germanium shows strong pinning S~ 0.005-0.002.
1.2.7.2 Contacts for Ge Transistors
Low sheet resistance metals are require for source and drain contacts in order to minimize series resistance in short channel devices. NiGe and PdGe are the best candidates for transition metal Germanides because they react at low temperature 150-360 0C. The drawback of these is degradation in morphology accompanied by increase in Rsh, which results in agglomeration after postgermanidation anneal at temperature higher than 400-500 0C. For NiGe the solution to overcome the problem is to introduce Yb which retards agglomeration and increase thermal stability. Also, NiGe and PdGe have high barrier for electrons and negligible for holes which makes it suitable for source and drain ohmic contacts for p-MOSFETs. Comparing the two, PdGe is less suitable because aqua- resia which etch unreacted Pd also etch Ge. The Ni can be etching with any acid compatible with Ge. So, this combined with low resistivity and accomodity price of Ni make NiGe is best candidate for integration [6].
1.2.7.3 Schottky Barrier Metal Source-Drain for Ge FETs
As devices scales down to small dimension, the unwanted series resistance, parasitic resistance are expected to degrade device performance. This occurs more when Ge is used as a channel because implanted Ge source and drain region suffers from enhanced diffusion, limited solubility, insufficient dopant activation all which contribute to parasitic, especially for n+/p Ge nMOSFETs, which are very challenging. The implant free SB metal source-drain Ge FETs may help in alleviating some of the problems reducing parasitic resistance and producing inherently ultra-shallow junctions suitable for scaled devices [6].
1.3 MOLYBDENUM DI SULPHIDE (MoS2)
MoS2 is a two dimensional material of the type MX2 where M is transition metal (Mo, W, Ru etc.) and X can be any chalcogen (S,Se,Te).It has an ultrathin layered structure and appreciable direct band gap of 1.9eV in monolayer regime. The two dimensional TMDCs have strong intralayer bonds and weak interlayer bonds that give rise to its layered structure and also to consequent anisotropic properties. MoS2 is the widely studied TDMCs because of its availability in nature in the form of molybdenite. TMDCs have been first studied since in 1960s in the field of dry lubrication, catalysis, photovoltaics and in batteries. MoS2 is the first one which shows its semi conducting properties [25].
The few layers MoS2 have good potential for application in optoelectronics, nanoelectronics and also in flexible devices. We have the capability for controlling its valley and spin degree of freedom which makes it suitable for spintronics and valleytronics application. The presence of the unsaturated d-orbital due to transition metal to material band structure introduces interesting properties for examples magnetism, charge density, waves and superconductivity that make them suitable for research. The atomically thin MoS2 flakes are the one which make them highly sensitive to environmental and substrate and also which affects every aspect of material properties such as the materials growth mechanism, its carrier transport processes, the performance of variety of few layer based MoS2 devices etc.
Fig1.3:MoS2 applications in various fields [24]
1.3.1 Crystal Structure
MoS2 has hexagonal structure in which both Mo and S are arranged hexagonally in a plane and Mo is sandwitched between two S-planes. It has many possible phases: 1T/2H/3R as shown in Fig 1.4. 1T MoS2 unit cell has octahedral shape and 2H unit cell has trigonal prismatic 3D view of these 1H/2T MoS2 structures along the top view of 2D planes is show in Fig 1.5. Top view of MoS2 and its unit cell configuration is shown in Fig 1.6. The thickness of each layer is 0.65nm in bulk MoS2 and these layers held together by weak Van der Wall forces. Mo-S bond length is 2.4”, lattice constant is 3.2” and separation between two S-planes is 3.1”. 2H are mostly used because of its semiconducting properties. 1T/3R is less stable and is not form in normal conditions. The commonly used configurations are hexagonal and octahedral. Bulk MoS2 belongs to D6h family and monolayer belongs to D3h space group of crystal structure. The compound which belongs to D6h shows inversion symmetry and which belongs to D3h does not show inversion symmetry [33].
Fig1.4: (a) The hexagonal structure of Mo and S layers, (b) Side view of 1T/2H/3R type structures of MoS2 [25]
Fig 1.5: Crystal structure of 2H and 1T phases, (a) Trigonal prismatic, (b) Octahedral [30]
Fig1.6: (a) Top view of hexagonal structure of monolayer MoS2, (b) Trigonal prismatic structures and octahedral structures of MoS2 [24]
Optical second harmonic spectroscopy suggests that MoS2 with odd number of layers does not possess inversion symmetry and with even number of layer possess inversion symmetry. MoS2 has very high Young’s modulus than steel. A monolayer of MoS2 is marginally stronger than bulk crystal. The MoS2 layer can be deformed upto 11% without occurrence of any fracture and can also be bent upto 0.75nm radius of curvature without losing its electronic properties [27]. These are the merits which make them good candidate for flexible electronics. It also has piezoelectric properties which can be used for formation of mechanical transducers. MoS2 resonators are operated on high frequency as well as on very high frequency and also shows very high figure of merit.
1.3.2 Band Structure
MoS2 and other TMDCs band structure depends on thickness. The band structure of bulk MoS2 is shown in Fig 1.7 along with the indication of crystal points. Bulk MoS2 is indirect band gap material 1.29eV and the conduction band minima is located halfway between ” and K and the valance band maxima is located at ” point which constitutes indirect band gap.
As the number of layers decreases the lowest band of the conduction band moves upward and increases the overall band gap of the MoS2. The state of conduction band at K-point is mainly due to the presence of d-orbital of Mo atoms and are also almost unaffected by interlayer interactions. The direct band gap of MoS2 at K-point is only increases by 0.05-0.1eV as shown in Fig 1.8(b). The states near ” point occur due to hybridization between pz – orbitals of sulphur atoms and d-orbital’s of Mo atoms and it is affected by interlayer interactions. So, as the number of layers decreases the bands at ” are more affected. In monolayer, the direct gap is smaller as compared to indirect gap and the value of smallest direct band gap is 1.9eV. There is a transition from indirect to direct band gap which show the inter layer interactions, quantum confinement and also long-range Coulomb effect.
Fig1.7: (a) Electronic Structure of bulk MoS2, (b) Brillouin zone and special points of hexagonal lattice system [27].
MoS2 commonly shows n-type behaviour. This is partially due to S vacancies in MoS2. These vacancies create deep midgap trap which can control n or p-type of behaviour of material. Also, if discuss about quantum confinement, MoS2 has large density of states which extends into the substrate and then final density is the total of density of states of substrate and that of MoS2. Due to large valance band and conduction band offset with MoS2, oxide substrate prevents any charge transfer between the two.
They argued that substrate impurities play an important role in Fermi level defining of MoS2. There are two types of impurities in SiO2: Na atoms and O-dangling bonds. They create donor and acceptor impurities respectively. The Fermi level of MoS2 aligns with that of impurity level. Fig 1.9 shows the effect of substrate impurities on MoS2. Band gap of MoS2 also changes with strain which can be used in sensors to monitor physical state of a system based on strain. There is a class of materials in electronics called straintronics whose electrical properties are strain dependent where MoS2 is widely used. Also, due to the dependence on spin and valley states at K-point in band
Fig1.8: Band structure of MoS2 (a) Direct and Indirect band gap of MoS2, (b) Transition of indirect to direct band gap as shown above from (a’d).
Fig1.9: (a) Schematic band diagram of MoS2 placed on the defect SiO2 substrate, (b) donor level, and (c) acceptor level [32].
1.3.3 Carriers
The effective mass of electrons (meff) in MoS2 is 0.48me at K-point which is much higher than graphene (0.012me). The interaction with acoustic phonons are used for determination of MoS2 carrier mobility at low temperatures opposed to optical phonons mobility limited at higher temperatures. The first principle study predicts the mobility to be 2450 cm2/Vs at lower temperature and 400 cm2/Vs at higher temperatures [24]. However, the experimental value of mobility is significantly lower. The devices made from MoS2 shows mobility between 1~20cm2/V-s.
Optical pumping method is used to increase the mobility and the formation of electron-hole pairs due to luminescence may propagate as a charge neutral pair and remains unaffected by Coulomb scattering and results in better mobility. The conductivity of MoS2 decreases under photoluminescence condition [26]; this is due to formation of stable trions at room temperature. These trions have equal charge to electrons/holes and increased effective mass results in decrease in conductivity.
1.3.4 MoS2 Fabrication
Andre and Kostya were the first successful in extracting very thin 2D sheets from bulk graphite. This is very simple in principle but effective method [24] is to be used since for extracting 2D sheets from layered materials. Other methods are growth by chemical reaction and mechanical exfoliation. These methods are described below:
(a) Mechanical Exfoliation
Mechanical exfoliation method is shown in Fig 1.10. In this, a pure bulk 2D material is peeled off from the surface by the use of adhesive tape. This process is repeated again and again until a very thin layer of 2D material is retained on the tape which is then transferred on the substrate by adhesion. This method is very simple and easy which makes it popular. The uniformity and yield are the main issues with this method. The size of MoS2 extracted by this method ranges from micrometers to few tens of micrometers [24]. Many solutions and modified methods for mechanical exfoliation have been experimented. If we use Pyrex glass as a substrate material with applied potential then this results in very thin MoS2 sheet on the surface [24]. The MoS2 observed under optical microscope needs to be good optical contrast with the substrate. This leads to minimum thickness of the substrate. SiO2 with minimum thickness 270-300nm is the mostly used substrate material.
(b) Chemical Exfoliation
Another method is chemical exfoliation which is known before mechanical exfoliation method. This method recently gained so much interest because it contains new physics and also applications of layered materials. Fig 1.11 shows the general flow for the growth of MoS2 followed by the transfer of these layers on the substrate. The growth of MoS2 is performed by chemical vapour deposition method (CVD) on a metal substrate which is transferred on the substrate by the use of sacrificial layer [24]. In general, there are two main chemical methods: intercalation method and solvent based method. Joenson et. al was the first who used the intercalation method in 1986 and it is also called Morrison method. These methods are described below:
Fig1.10 : Mechanical exfoliation of MoS2 2D crystals. (a) Adhesive tape is pressed against a 2D crystal so that top few layers are attached to the tape (b), (c) The tape with crystal of layered material is pressed against the surface of substrate of choice, (d) Upon peeling off, and bottom layer is left on the substrate [34].
(i) Intercalation Method
Intercalation of Li ions inside the vdW gap between 2D monolayers can be used to separate two layers. The amount of intercalation is not controlled very precisely which results in lower yield. This method requires high temperature and take days to complete the process. A modification in this method is proposed by Zeng et.al. is the formation of using electrochemical process instead of using pure chemical process to control the amount of intercalation. They reported in 92% yield of monolayer MoS2 flakes. Also, high yield of 2H phase is observed which was problem with simple intercalation based method where 2H phase is converted into 1T metallic phase.
(ii) Solvent-based Exfoliation Method
This method is also called Coleman method which is first studied in 2011 and is comparatively new method. In this, we use the organic solvent to extract monolayer or few layers MoS2 from bulk by reducing the energy required for separation. This method retains 2H- phase of MoS2 which increase the yield. Sonication is needed to separate MoS2 in the solvent. The dimension reduces to few hundred nanometres due to high energy sonication [24]. Larze size flakes are required in wafer scale to fabricate chips in bulk amount. Hence, the full potential of MoS2 cannot be used until technology for large scale uniform and controllable growth of MoS2 should be developed.
Fig1.11: Transfer of CVD grown on 2D crystals. (a),(b) 2D crystals are grown by CVD on a surface of metal, (c) A sacrificial layer is deposited on the top of 2D crystals, (d) The metal is etched away, leaving 2D crystal stuck away on the sacrificial layer, (e)the sacrificial layer with 2D crystal is transferred on the substrate of choice,(f) The sacrificial layer is removed [34].
(iii) Other Methods
Despite of few layer of MoS2 have unique and interesting properties but their full potential cannot obtained if it is not grown at wafer scale. The exfoliation methods used have two problems one is a flake size and other is yield limitation so they are not used for integrated circuit fabrication and for other advanced uses. Till now, thin MoS2 could be first grown using these three precursors: ammonium thiomolybdate [(NH4)2MoS4] solution, a thin deposition of elemental molybdenum or molybdenum trioxide [MoO3]. The first precursor ammonium thiomolybdate which can be dissolved in dimethylflormamide and then used in form of solution and after this the solution is either coated on a substrate or used in vapour form through the use of bubbler as shown in Fig 1.12. Liu et al. dipped their substrate in ammonium thiomolybdate solution and after this pulled out their substrate vertically at a constant rate in order so that they obtain a uniform thin film. After this, the substrate is annealed in Ar/H2 at 500 0C and low pressure 1 torr and then second anneal at 1000 0C at a pressure of 500 torr was followed. At the time of anneal, there is a decomposition of thiomolybdate occurs which form MoS2, H2S, S and NH3 gases. MoS2 which is grown on sapphire substrate shows larger grain and better quality as compared to those which are grown on SiO2 substrate. The thickness of layer which is grown is about 2-5nm and it also depends on precursor dose [24].
Fig1.12: Growth technique of MoS2 using ammonium thiomolybdate [24].
The second precursor method consists of a thin layer of molybdenum which is first evaporated onto substrate and after this it is followed by sulfurization at high temperature and also under inert conditions as in Fig 1.13. At high temperature, more crystalline MoS2 is grown which covered the entire substrate.
Figure1.13 Growth technique of MoS2 using elemental molybdenum
Fig1.14: Growth technique of MoS2 using molybdenum trioxide [24].
The third choice of precursor is MoO3. Balendhran et al. were the one who first use this technique. They have annealed their substrate in the presence of powdered MoO3 and kept in crucible and found that good quality layer is obtained after the annealing at 830 0C for 180 min as shown in Fig 1.14 [35].
1.3.5 Raman and PL characterization
Raman spectroscopy is mostly used to characterize the materials and also it inspects the crystal quality as well as number of layers in MoS2. It has four Raman-active modes (E1g,E2g1,E2g2 ,A1g) and two IR-active modes (A2u and E1u). These are shown in Fig 1.15. E2g1 and A1g undergo red and blue shift respectively with increasing number of layers. The former depends on the long chain columbic interactions and latter depends on the product of interlayer interactions [24].
There also exists E2g2,B2g2 modes. These are low frequency modes called shear and breathing modes respectively as shown in Fig 1.16 (a), (b). These modes shows energy shift also based on applied strain. Fig 1.16(c) shows the comparison of mono, bi-, tri- layers of MoS2 profile.
PL spectra are used for characterization of monolayer properties. Very high yield (>104) of PL intensity in monolayer MoS2 compared to bulk which supports direct band gap in monolayer MoS2 argument. The reduction of monolayer MoS2 band gap can be seen in PL spectra. The result from PL spectra is shown in Fig 1.17. Photoluminescence (PL) test can be used with Raman in characterization and determining the quality of MoS2 growth especially for monolayer growth quality and crystallinity [36]. Direct band in bulk MoS2 also exists, PL in bulk is nonexistence due to excitonic absorption but in monolayer direct band gap is the principle method for direct band radiative recombination is the dominant method for excitonic recombination.
Fig1.15: Raman active and two IR-active modes [27].
Fig1.16: (a) Breathing modes, (b) Shear modes, (c) Raman spectra for mono, bi-, tri- layer of MoS2 [28].