MoS2 Thin Films for Photo-Voltaic Applications

The low dimensional chalcogenide materials with high band gap of ~1.8 eV, specially molybdenum di-sulfide (MoS2), have been brought much attention in the material science community for their usage as semiconducting materials to fabricate low scaled electronic devices with high throughput and reliability, this includes also photovoltaic applications. In this chapter, experimental data for MoS2 material towards developing the next generation of high-efficiency solar cells is presented, which includes fabrication of ~100 nm homogeneous thin film over silicon di-oxide (SiO2) by using radio frequency sputtering at 275 W at high vacuum~10−9 from commercial MoS2 99.9% purity target. The films were studied by means of scanning and transmission electron microscopy with energy disperse spectroscopy, grazing incident low angle x-ray scattering, Raman spectroscopy, atomic force microscopy, atom probe tomography, electrical transport using four-point probe resistivity measurement as well mechanical properties utilizing nano-indentation with continuous stiffness mode (CSM) approach. The experimental results indicate a vertical growth direction at (101)-MoS2 crystallites with stacking values of 7-laminates along the (002)-basal plane; principal Raman vibrations at E2g at 378 cm−1 and Ag at 407 cm−1. The hardness and elastic modulus values of H = 10.5 ± 0.1 GPa and E = 136 ± 2 GPa were estimated by CSM method from 0 to 90 nm of indenter penetration; as well transport measurements from −3.5 V to +3.5 V indicating linear Ohmic behavior.


Introduction
Layered chalcogenide materials have been of high relevance since almost 40 years for their diverse applications such as tribology [1], chemical catalysis [2] and nowadays as semiconductors towards development of high-throughput and energy efficient transistors and devices [3,4]. MoS 2 is a two-dimensional material with a band gap ranging between 0.9 and 1.8 eV as calculated theoretically by first principles methods and as measured experimentally by Kam & Parkinson using photo-spectroscopy as a function of crystal orientation [5,6]. The crystal structure of MoS 2 is hexagonal with space group R3m (a = b = 3.16 Å and c = 18.41 Å), having d-bonded layers of S-Mo-S along a-b plane which are stacked along c-axis by weak Van der Waals forces with 6.2 Å of separation within layers [7]. The crystal structure was studied using electron microscopy techniques as described by Chianelli et al. who were able to observe its layered structure [8]. However, electron beam dosage during electron microscopy studies plays an important role to avoid any structural damage as described by Ponce et al. when using TEM technique who concluded high-resolution imaging at operational voltages near ~80 kV [9] to be possible. By in-situ TEM, Helveg et al. were able to synthesize small clusters of MoS 2 from molybdenum oxide and hydrogen sulfide gases at beam radiation dosage of 100 e − /Å 2 s [10]. The mechanical properties were studied by Casillas et al. achieving an atomistic observation of a resilient nature on MoS 2 laminates at 8GPa of external applied pressure and its mechanical recovery during in-situ AFM on TEM sample holder [11]. Applying atomic force microscopy (AFM), Bertolazzi et al. determined a Young modulus values of 270GPa ± 100GPa and fracture strength of 16~30GPa in MoS 2 layers as suspended in patterned silicon wafers [12,14], and Castellanos-Gomez et al. estimated an average Young modulus E = 330GPa in suspended MoS 2 sheets over patterned silicon wafer [13]. The mechanical properties were studied by density functional theory and molecular dynamics, Jiang et al. calculated a theoretical Poisson's ratio value of v = 0.29 applying Stillinger-Weber potential [15]. The reactive empirical bond-order (REBO) potential was used by Li et al. to understand structural effects at chemical bonding within S-Mo-S layers, their findings indicate induced vacancies on the basal plane can influence Poisson's ratio values [16]. The atom probe tomography enables the chemical understanding with three-dimensional spatial resolution and was applied to determine dopants, contamination and ionic distribution within semiconducting matrix [17], Singh et al. used APT technique to determine distribution of Ti over MoS 2 matrix [18]. Regarding electrical transport, Lia et al. [4] and Samuel et al. [38] performed transport electrical measurements encountering a linear ohmic behavior in MoS 2 . This chapter covers mechanical, electrical and microstructure characterization by electron microscopy, low angle x-ray, atom probe tomography and CSM-nanoindentation to obtain information about crystal growth, elastic modulus (E), hardness (H) and electrical transport on MoS 2 films.

RF sputtering
The Molybdenum di-Sulfide (MoS 2 ) films were fabricated with a high vacuum Kurt J. Lesker© PVD 75 machine; applying RF-sputtering at a rate of 2.26 Å/sec at 275 W of plasma power over 4″-diameter silicon oxide (SiO 2 ) wafers. The films were deposit from commercial MoS 2 99.9% targets (Kurt J. Lesker). By using dwell time of 300 seconds a film thickness value of ~100 nm was achieved as indicated by profilometry measurements, Figure 1E.

Scanning electron microscopy
The film morphology and crystallographic structure were investigated using scanning and high-resolution transmission electron microscopy (SEM, TEM). SEM was performed in a Hitachi® SU5500 unit, equipped with Energy-dispersive X-ray spectroscopy (EDS) unit and operated at 30 kV with 8A of current to avoid surface damage on the film. Observations indicate a high-degree of porosity and vertically aligned MoS 2 film matrix, as presented in Figure 2A-C which is in agreement with Kong et al. [18]. EDS analysis reveals the two main signals that correspond to Sulfur-K α and Molybdenum-L α at 2.4 keV, as presented in Figure 2E, in agreement with Lince & Fleischauer [20].

Transmission electron microscopy and atom probe tomography
The microstructure of MoS 2 thin matrix was also studied using Scanning Electron Transmission Microscopy (STEM) using a Cs-corrected 2200-JEOL, with STEM unit, equipped with a high-angle annular dark-field (HAADF) detector, X-Twin lenses and CCD camera. A lamella was prepared using Focus-Ion Beam model JEOL JEM 9320 at 30 kV and 25 mA, MoS 2 film surface coated with gold and gallium. Atom Probe Tomography (APT) was performed on Cameca® LEAP 4000X high-resolution system in laser pulse mode (wavelength ~355 nm), measurements were taken at 60 K with evaporation rate of 0.5 and laser frequency of 100 kHz, laser beam was set to 70pJ/V, all data was reconstructed using IVAS© 3.6.10a package. The samples were prepared using focus ion beam FEI Strata dual-beam instrument coupled with micromanipulator Oxford® Omniprobe® 200 by lift-out method as described by Szász et al. [21]. MoS 2 film surface was protected using platinum layer, and cuts were done at 30 kV at 260pA gun power. Using both techniques, it was possible to determine chemical composition, and spatial resolution of S-Mo-S distribution along film matrix, stacking and orientation, as presented in Figure 3. In the image the top part corresponds to MoS 2 and uppermost bright layer is due to gold gallium coating. The atom probe tomography (APT) is a technique used to understand in a threedimensional reconstruction with high-spatial resolution the chemical distribution and composition as indicated by Kelly & Miller [17]. A sample is placed in the main APT chamber to undergo an ionizing evaporation process at a high electric field triggered by a laser pulse; the potential energy of an atom at the sample surface, as caused by the applied voltage on the sample neV, is converted into kinetic energy ~1/2mv 2 in the vicinity of the tip. This relationship, in order to understand the mass-to charge-state ratio m/n of evaporated ions, is given by Eq. (1); with n as number of electrons removed from the ion, e electron charge (−1.62 × 10 −19 C), V total applied voltage, m is atomic mass and speed of atoms are given by conventional v = d/t, which is with good approximation constant, distance d and lastly t is the time of flight, as described by the schematic drawing taken from Kelly and Larson [23]. Short laser pulses (<1 ns) are used for APT and can field evaporated for almost any material regardless of its electrical conductivity as described by Kellogg et al. [24]. (1) Nowadays, usage of APT to survey spatial distribution of atomistic species in semiconducting devices like n-doped metal-oxide field effect transistors [25] and Singh et al. applied with high success to titanium-MoS 2 and strontium oxide-MoS 2 films [18]. In this case, APT measurements were performed to understand the spatial distribution of MoS 2 film matrix. Figure 4 illustrates the preparation of APT samples using a FIB (Figures 5-9).

Raman spectroscopy
The Raman spectroscopy was obtained using Alpha 300RA system equipped with a 532 nm Nd-YAG laser and a 100X 0.9 NA objective. The laser power was varied to avoid surface damage; with no additional sample preparation during study. Modes of vibration at E 1 2g = 378 cm −1 and A 1 g = 407 cm −1 are indicators of sulfur vibrations caused by dangling bongs on S-Mo-S chemical structure as indicated schematically Figure 10 (insets).

Grazing incidence X-ray diffraction (GIXD)
X-ray diffraction was collected using a Panalytical X-Pert system with source of Cu Kα λ = 1.41 Å radiation. The grazing incidence angle was fixed at 0.5 o with 20 o < θ < 80 o and step size of 0.02 o with a graphite flat crystal monochromator, described by Liu et al. while characterizing same layers of MoS 2 [26] and presented in Figure 11.

Nanoscale mechanical properties
The nanoscale mechanical properties were evaluated to obtain Elastic modulus (E) and Hardness (H) of MoS 2 thin films; this was possible using an Agilent nanoindenter model G200 coupled with a DCM II head instrument and Berkovich diamond indenter tip radius of 20 ± 5 nm, penetration depth limit of 400 nm, strain rate of 0.05 s −1 , and harmonic displacement and frequency of 1 nm and 75 Hz, Poisson's coefficient of ν = 0.22. The equipment was calibrated using a standard  [27]. All data was recorded by AFM Nano Vision© system attached to the nanoindenter system. The estimated values for hardness (H) and elastic modulus (E) were calculated using Eq. (2) to   determine stiffness S, when comparing to silicon substrates (001) surface termination and applying a continuous stiffness method as described extensively by Pharr et al. [28].
In Eq. (2), ω is the excitation frequency, (Zo) displacement amplitude, (φ) phase angle, and (F0) is the excitation amplitude, all those values can be obtained if the machine parameters load-frame stiffness Kf and stiffness of springs (Ks) as well the mass m are known input values during nanoindentation test. The coating hardness of film H f can be estimated using a work indentation model described by

2D Materials
Eq. (3), having Hc as composed film/substrate hardness; Hs and Hf as substrate and film hardness, the constant k represents a fitting parameter determined experimentally from the variation of Hc with relative indentation depth (β = Hc/t). (3) The elastic modulus E can be estimated using Eq. (4); having Eeff as effective reduced elastic modulus of the system in array film/substrate, contact area is determinate by A as function of the penetration depth, ν is the Poisson ratio, t represents film thickness, and α is a parameter which depends on the material and the indenter   geometry, in our case a pyramidal shape, as described by Domínguez-Rios et al. [29] and Hurtado-Macias et al. [30]. (4) By using CSM method, it is was possible to estimate elastic modulus and hardness values as follows: Three regions of test are observed in the Figures 12 and 13, where region I is hardness values for MoS 2 crystallites with penetration depth of 0-90 nm, having no influence from silicon oxide substrate and a hardness value of H = 6.0 ± 0.1 GPa and elastic modulus of E = 136 ± 2 GPa. The region II, which has a penetration deep of 90-120 nm both values of elastic modulus and hardness are increased, meaning a clear influence by silicon oxide substrate, as confirmed by profilometry a thin film thickness of ~105 nm (both insets of Figure 1). The region III with penetration depth of 120-150 nm represents a hardness and elastic modulus of silicon oxide substrate, which are in partial agreement with Malzbender & With [31] to whom performed similar experiment on SiO 2 spin coated with methyltrimethoxysilane. Figure 10. The Raman spectra with two characteristic modes of vibrations at E 1 2g at 378 cm −1 and A 1 g at 407 cm −1 , in agreement with Kong et al. [19].

Figure 11.
Grazing incidence x-ray diffraction it was possible to observe a dominant (101) reflection at 2θ ~30°, in agreement with Liu et al. [26] for vertical aligned layers.
The obtained values for hardness and elastic modulus are smaller estimations when comparing with results as presented by Bertolazzi et al. [12,14] for single layers of MoS 2 ; we believe this occurs because of low dimension laminates can be stronger than stacking of MoS 2 crystallites. The applied force was done over (001)-basal plane as suspended on patterned silicon holes [12,14], and in this case indenter tip can sweep MoS 2 crystallites over surface area. For that reason, our research team proceed to estimate film adherence by using AFM scratching technique in Figure 12. Nanoindentation curves estimated experimentally using continuous stiffness method (CSM), the curve corresponds to regions I, II, III. In region I the estimated elastic modulus is E = 136 ± 2 GPa corresponds to 0-90 nm of penetration depth, which is indicated to be only for MoS 2 film, in agreement with [28].  encountering a deformation 0.85 μm 2 with a residual groove width 1 μm (a total groove height 125 nm and pile up height 40 nm), as presented in Figure 14, along with indentation sites completed to obtain elastic modulus and hardness values.

Electrical transport and resistivity
The electrical transport of the MoS 2 film matrix was investigated using fourpoint probe method as indicated in Figure 15, equipped with Keithley 4200-SCS in applied voltage range from −3.5 to 3.5 V. The transport measurements were done at room temperature and by direct contact to the MoS 2 film surface, no especial solder or metallic glue was used. Also, they were completed in the presence of light and dark conditions, the results indicate a linear Ohmic behavior, as presented in

Conflict of interest
Authors declare no conflict of interest.