Temperature-Dependent Morphology Modulation of MoO₂ from 1D Nanoribbons to 2D Nanoflakes for Enhanced Two-Dimensional Electrode Applications

Abstract

The morphology modulation of target crystals is important for understanding their growth mechanisms and potential applications. Herein, we report a convenient method for modulating the morphology of MoO₂ by controlling different growth temperatures. With an increase in growth temperature, the morphology of MoO₂ changes from a nanoribbon to a nanoflake. Various characterization methods, including optical microscopy, atomic force microscopy, (vertical and tilted) scanning electron microscopy, Raman spectroscopy, high-resolution transmission electron microscopy, and selected area electron diffraction, were performed to unveil the morphology modulation and lattice structure of MoO₂. Both MoO₂ nanoribbons and nanoflakes display a standing-up growth mode on c-sapphire substrates, and their basal planes are MoO₂(100). Further investigations into devices based on MoS₂ with Au/Ti/MoO₂ electrodes show the potential applications of MoO₂ in two-dimensional electrodes. These findings are helpful for the synthesis of MoO₂ with different morphologies and applications in the field of optoelectronic nanodevices.

1. Introduction

Molybdenum oxides have been extensively studied due to their distinct physicochemical properties and wide range of applications in photodetectors, photocatalysts, electronics, and organic photovoltaic devices [1,2,3,4,5]. Among them, molybdenum trioxide (MoO₃) and molybdenum dioxide (MoO₂) have attracted much more research interest [6,7]. While MoO₃ exhibits the features of an n-type semiconductor with a large bandgap (~3.2 eV) [5,7], MoO₂ exhibits the features of a metal with a monoclinic structure and lattice parameters of a = 5.62 Å, b = 4.86 Å, c = 5.63 Å, and β = 120.93° (ICSD-152316) and possesses a high melting point and excellent chemical stability [8,9,10]. The high electronic conductivity of MoO₂ indicates its potential as a promising anode material for next-generation lithium-ion batteries [11]. Its low-symmetry structure endows MoO₂ with highly anisotropic electrical and optical properties [12,13]. In addition, MoO₂ nanosheets show linear magnetoresistance [9,14], showing potential for applications in new magnetic devices.

Due to its non-layered structure, the morphology of MoO₂ can be modulated by controlling different growth conditions, such as substrate and precursor density [6,8,12,15,16]. In fact, different morphologies of MoO₂ have been achieved over the past decades. For instance, 1D MoO₂ nanorods were prepared via a thermal decomposition method, showing great capacitive behavior [17]. Chemical vapor deposition (CVD) is one of the most popular and controllable synthesis methods for producing high-quality 2D materials. Various factors, including precursors, gas flow, substrate, and growth temperature (GT), can influence the structure, composition, and morphology of target crystals [18,19,20,21]. MoO₂ nanosheets on SiO₂/Si substrates synthesized via the CVD method exhibited great in-plane anisotropy at an electrical conductivity ratio of about 10.1, which demonstrated potential for application in multifunctional integrated plasmonics and ion-inspired electronics devices [12]. Recently, MoO₂ nanoflakes were investigated in the lying-down and standing-up growth modes on c-sapphire substrates, and the products showed morphological competition during the growth process [8]. Consequently, the morphology modulation of MoO₂ is essential for understanding their growth mechanism, as well as exploring their applications in different areas.

In this article, MoO₂ was synthesized through the atmospheric-pressure chemical vapor deposition (APCVD) method, using MoO₃ powder and H₂ as the precursor and reductant, respectively. The effects of GT on MoO₂ morphology were systematically investigated. The morphology of as-prepared MoO₂ crystals changed from a nanoribbon to a nanoflake with an increase in GT. Comprehensive characterization techniques, including optical microscopy (OM), atomic force microscopy (AFM), scanning electron microscopy (SEM), Raman spectroscopy, high-resolution transmission electron microscopy (HRTEM), and selected area electron diffraction (SAED), were performed to elucidate the morphology modulation and lattice structure of MoO₂. The basal planes of both MoO₂ nanoribbons and nanoflakes were unveiled to be MoO₂(100), and the growth direction of MoO₂ nanoribbons was MoO₂<010>. Energy-dispersive X-ray spectrometry (EDX) measurements showed a stoichiometric ratio of ~2.11 between the Mo and O elements in an as-prepared sample, a result that was consistent with the composition of MoO₂. Due to its high electronic conductivity, MoO₂ can be used as an electrode material for nanodevices. Field-effect transistor (FET) devices based on MoS₂ nanoflakes were then fabricated with Au/Ti and Au/Ti/MoO₂ electrodes, showing better performance for Au/Ti/MoO₂ electrodes. These findings offer valuable insights into the synthesis of MoO₂ with different morphologies and show their potential for application in two-dimensional (2D) electrode and optoelectronic nanodevices.

2. Materials and Methods

2.1. Sample Synthesis Method

The MoO₂ crystals were prepared in an APCVD system, as shown in Supplementary Figure S1a, with one heating zone. MoO₃ powder (99.95%, Aladdin, Shanghai, China) was treated as the metal precursor, and hydrogen as the reductant. Before conducting the growth experiments, acetone, isopropanol, and DI water were used to preclean the c-sapphire substrates. During the growth process, the c-sapphire substrates were supported with a SiO₂/Si sheet and placed above the metal precursor within a quartz boat, which was placed in the center of the heating zone. The APCVD system was initially heated from room temperature to 150 °C, with an argon flow rate of 200 sccm. After holding the temperature at 150 °C for 10 min, the APCVD system was heated to GT for 40 min. When the temperature reached GT, hydrogen was introduced at 0.5 sccm, and the argon flow changed to 40 sccm. The growth process lasted for 20 min. After the growth process ended, the system temperature was dropped to 600 °C, followed by rapid cooling at an argon flow of 200 sccm. The temperature–time profile for the growth experiment is shown in Supplementary Figure S1b.

2.2. Device Fabrication

The MoS₂ nanoflakes were transferred onto SiO₂/Si substrates via the PMMA-assisted method, and the MoO₂ nanoflakes were transferred onto MoS₂ nanoflakes via the PDMS-assisted method, as shown in Supplementary Figure S2. The photoresist (AZ 5214-E, AZ Electronic Materials, Shizuoka, Japan) was coated onto a MoO₂/MoS₂ heterostructure at a 4000 rpm spin-coating rate for 60 s and then baked at 105 °C for 1 min. The electrode patterns were exposed using a laser writer (MicroWriter ML 3, Durham Magneto Optics Ltd., London, UK). The exposed part of the photoresist was removed via AZ726 MIF (AZ Electronic Materials, Shizuoka, Japan) for 30 s. Additionally, the developer was removed by floating the photoresist in DI water for 30 s. Finally, 10 nm/50 nm Ti/Au metals were deposited onto the samples to fabricate the electrodes through e-beam deposition.

2.3. Sample Characterization

The OM measurements were performed via Leica optical microscopy (Leica DM2700M RL, Wetzlar, Germany). The Raman spectra were selected using a Raman microscope (WITec alpha 300R, Lise-Meitner-Str., 6 D-89081 Ulm, Germany) under a 532 nm laser with 600 lines/mm grating. The SEM measurements were carried out using a Zeiss Sigma HD scanning electron microscope (Carl Zeiss AG, Oberkochen, Germany). The AFM measurements were performed under ambient conditions via a Bruker Dimension Icon AFM (365 Boston Rd., Billerica, MA, USA). The TEM, EDX, and SAED measurements were conducted in a JEM-3200FS transmission electron microscope (JEM-3200FS, JEOL, Street No. 6, Haidian District, Beijing, China) using an acceleration voltage of 300 kV.

3. Results and Discussion

Figure 1(a₁–a₆) displays OM images of MoO₂ on c-sapphire substrates grown at different GTs. When the GT is 750 °C, only particles can be observed on the substrate, indicating that MoO₂ nanoribbons and nanoflakes cannot be achieved at a growth temperature of 750 °C. After raising the GT by 10 °C, MoO₂ nanoribbons tens of micrometers in length were achieved on the substrate. Although these samples looked like nanorods [6], they exhibited a standing-up growth mode. As a result, correctly evaluating the width of the MoO₂ nanoribbons was difficult. With a continuous increase in GT, the length of the MoO₂ nanoribbons became shorter. Furthermore, the products displayed an obvious defocus, indicating that MoO₂ grew along the out-plane direction. When the GT was 800 °C, standing-up MoO₂ nanoflakes predominantly grew on the substrate. To study the modulation of GT on the morphology of MoO₂, MoO₂ samples grown at different temperatures were transferred onto silicon substrates. As shown in Supplementary Figure S3, it was observed that, by raising the GT to 800 °C, MoO₂ transformed from a 1D nanoribbon to a 2D nanoflake gradually. In fact, due to the standing-up growth mode, the width of the MoO₂ exhibited significant changes even though its morphology transformed from that of a 1D nanoribbon to a 2D nanoflake. In addition to the standing-up samples, a few lying-down MoO₂ nanoflakes in the shapes of triangles and isosceles trapezoids were observed on the substrate, which may be due to the inevitable vibration during the sample transfer process.

To unveil the growth behavior of MoO₂ with different morphologies, SEM measurements were further performed. Figure 1b,c show the SEM images taken along the out-plane direction. As shown in Figure 1b, MoO₂ grown at lower GTs clearly exhibited a nanoribbon morphology. The unique wavy texture along the growth direction suggests that the grown MoO₂ nanoribbons are flexible. It is worth noting that a shaded area can be observed on one side of the MoO₂ nanoribbon (marked by a dashed red rectangle), indicating the standing-up growth mode of the nanoribbons. At higher GTs, the grown MoO₂ nanoflakes displayed shorter lengths and wider widths, as shown in Figure 1c. Consistent with the nanoribbon sample, a shaded area also appeared on one side of the MoO₂ nanoflakes due to the standing-up growth mode. Figure 1d,e present the SEM images of MoO₂ nanoribbons and nanoflakes taken at rotation angles of 30°, respectively, which clearly demonstrates the standing-up growth mode of MoO₂. The defocus phenomenon in the rotated SEM images is due to over-focusing and under-focusing during rotation-angle induction. A close-up SEM image of the MoO₂ nanoribbon, as shown in Figure 1f, indicates that the basal plane of the nanoribbon formed an included angle with the substrate surface, resulting in the presence of the shaded area in Figure 1b. A wavy structure was seen on the top side of the MoO₂ nanoribbon, which may be due to the residual strain induced by lattice mismatch between the MoO₂ and substrate. The MoO₂ nanoflakes also displayed a standing-up growth mode, and their morphologies were in the shapes of triangles and right-angled trapezoids, as shown in Figure 1g. In fact, the MoO₂ nanoflakes exhibited the shape of right-angled trapezoids instead of isosceles trapezoids, arising from the rotation of the MoO₂ sample. The above results clearly prove the existence of the standing-up growth mode and morphology change from a 1D nanoribbon to a 2D nanoflake of MoO₂ under different GTs.

Figure 2a,b display OM images of MoO₂ transferred onto SiO₂/Si substrates grown at different GTs, showing the morphology transformation from a 1D nanoribbon to a 2D nanoflake. Figure 2c,d present corresponding AFM images of the MoO₂ nanoribbon and nanoflake, in which the thicknesses were determined to be 9 and 8 nm, respectively, according to their height profiles. The Raman spectrum of the MoO₂ nanoribbon (nanoflake), as shown in Figure 2e and Supplementary Table S1, displays Raman peaks at 121.6 (126.4), 205.0 (207.4), 228.7 (228.7), 348.4 (346.1), 362.4 (364.7), 498.7 (498.7), 571.7 (573.9), and 744.9 (747.1) cm⁻¹, results that are in agreement with those of previous reports [5,9,14]. Figure 2f shows the Raman mapping images of the MoO₂ nanoribbon and nanoflake at 745 cm⁻¹, indicating the uniform quality of these samples. In addition, Supplementary Figure S4a exhibits the XRD patterns of the MoO₂ nanoribbon and nanoflake, suggesting the formation of high-quality MoO₂ phase. Supplementary Figure S4b presents the XPS spectra of the MoO₂ nanoribbon and nanoflake. According to the Mo 3d core-level emission, the two major peaks located at around 232.3 (Mo 3d_3/2) and 229.2 eV (Mo 3d_5/2) indicate the presence of Mo (+4). The XPS results are consistent with the values for MoO₂ reported previously [10].

To investigate the microstructure of the as-grown MoO₂ samples, TEM and SAED measurements were performed, as shown in Figure 3. The MoO₂ nanoribbon and nanoflake were transferred onto a holey carbon grid through a simple immersion and scraping method. Figure 3a displays a low-magnification TEM image of the MoO₂ nanoribbon. The red dashed line in Figure 3a represents the side of the MoO₂ nanoribbon adjacent to the substrate. The HRTEM image, as shown in Figure 3b, was selected in the white rectangle area in Figure 3a. The spacing between the two planes was determined to be 0.245 and 0.278 nm, corresponding to the (020) and (10−2) planes of MoO₂, respectively [6,14]. Figure 3c presents the corresponding SAED pattern of the MoO₂ nanoribbon. The real distances between two spaces, ~0.483 and 0.276 nm, were calculated based on the diffraction spots, corresponding to the (010) and (10−2) planes of MoO₂, respectively [6,14]. A low-magnification (LM) TEM image of the MoO₂ nanoflake is presented in Figure 3d, showing a similar contour to that in the OM results. Figure 3e,f display the HRTEM and SAED results of the MoO₂ nanoflake, similar to that of the MoO₂ nanoribbon, indicating that controlling GT only modulates the morphology of the target MoO₂, rather than the crystal lattice. Figure 3g shows the dark-field TEM image of an individual MoO₂ nanoflake. The corresponding EDX mapping images are presented in Figure 3h,i, showing uniform chemical composition in MoO₂. Figure 3j–l present the corresponding results for the MoO₂ nanoribbon. Supplementary Figure S5 displays the EDX spectrum of the as-prepared sample, showing a 1:2.11 stoichiometric ratio between the Mo and O elements, confirming that the as-grown products were MoO₂.

According to the I–V curves of the as-prepared MoO₂ nanoribbon and nanoflake in Supplementary Figure S6a,b, their electric conductivities were 5.6 × 10⁴ and 2.2 × 10⁴ S/cm at room temperature, respectively. Supplementary Table S2 lists the performance comparison among the MoO₂ nanoribbon and nanoflake with previous reports, confirming the high electric conductivity of the as-synthesized samples. The MoO₂ nanoflake is able to sustain stability until it approaches its electrical breakdown point at a V_DS of 6 V and a J_DS of 6.4 × 10⁷ A/cm², as shown in Supplementary Figure S6c. Due to the atomic thickness of 2D materials, such as single-layered MoS₂, the Fermi-level pinning effect is present in the interfaces between 2D materials and metal electrodes, arising from several factors, such as the defects created by the high-energy metal deposition and lithography process [22]. Therefore, MoO₂ nanoflakes can be employed as a 2D electrode material for devices based on ultrathin 2D materials to reduce the surface states from defects or residues at the contact point between metals and semiconductors. The AFM image of device and the typical Raman spectra of MoS₂ and MoO₂/MoS₂ are displayed in Supplementary Figure S7, indicating that the MoS₂ domain on the substrate was monolayer [23,24,25]. The as-prepared MoO₂ nanoflakes are transferred onto a monolayer MoS₂ to protect the MoS₂. Figure 4a,b show a schematic diagram and an OM image of the fabricated device, respectively. One pair of electrodes (electrodes 1 and 2, E₁₂) was deposited on the MoO₂/MoS₂ area as drain and source electrodes. For comparison, another pair of electrodes (electrodes 3 and 4, E₃₄) was directly deposited on an individual MoS₂ domain. Figure 4c presents the transfer curves of a MoS₂ device fabricated with E₁₂ and E₃₄ electrodes, showing the n-type transport feature [26,27,28]. The source–drain current of the MoS₂ channel with an E₁₂ electrode was one order of magnitude higher than that with an E₃₄ electrode under V_G = 40 V, suggesting better interface condition between MoO₂ and MoS₂. The electron mobility (μ) of the device was calculated from the transfer curves according to the following equation:

$$\mu ={\displaystyle \frac{L}{W}}{\displaystyle \frac{d}{{\epsilon }_0{\epsilon }_r}}{\displaystyle \frac{1}{V_{DS}}}{\displaystyle \frac{d{I}_{DS}}{d{V}_G}},$$

(1)

where L and W are the channel length and width of the device, ε₀ = 8.85 × 10⁻¹² F·m⁻¹ is the vacuum permittivity, ε_r = 3.9 is the relative permittivity of the SiO₂ dielectric layer, and d = 300 nm is the thickness of the SiO₂ layer. The electron mobility of the MoS₂ device is 9.04 cm²/V·s with Au/Ti/MoO₂ electrodes and 0.284 cm²/V·s with Au electrodes. Figure 4d,e display the output curves of the MoS₂ device fabricated with E₁₂ and E₃₄ electrodes under different V_G values, respectively, further suggesting better electrical performance with the Au/Ti/MoO₂ electrodes. The yellow arrows in Figure 4d,e represent the increasing trend of V_G value. Figure 4f,g present the time-dependent photocurrent measurements of a photodetector based on MoS₂ with Au/Ti/MoO₂ and Au/Ti electrodes and illuminated using 470, 532, 633, and 690 nm lasers. To evaluate the photoresponse performance, the device’s figures of merit, including photoresponsivity (R), specific detectivity (D*), and external quantum efficiency (EQE), were calculated by the following equations:

$$R={\displaystyle \frac{I_{ph}}{PS}},$$

(2)

$$D^{\ast }=R\sqrt{{\displaystyle \frac{S}{2q{I}_{dark}}}},$$

(3)

$$EQE={\displaystyle \frac{R}{(hc/e\lambda )}},$$

(4)

where P is the power density, S is the effective area of the device, h is the Planck constant, and c is the speed of light [29,30,31]. It should be noted that the effective area, S, was defined as the entire MoS₂ due to the fact that it was not patterned. Therefore, the calculated photoresponse performance was conservative and lower than the actual value according to the equations. Supplementary Figure S8 plots the evolution of the above figures of merit under different power densities at 532 nm laser illumination. In particular, the device fabricated with Au/Ti/MoO₂ electrodes exhibited a maximum R value of 9.0 A·W⁻¹ at a low light power density of 0.05 mW·cm⁻². Meanwhile, the device exhibited the highest D* and EQE values of 3.5 × 10¹¹ Jones and 2028%, respectively. For the sake of comparison, Supplementary Figure S9 compares the photoresponse performance between the MoS₂ devices fabricated with and without the MoO₂ layer, demonstrating more efficient photoresponse for the MoS₂ device with Au/Ti/MoO₂ electrode. These experimental results indicate that the introduction of MoO₂ between a metal electrode and MoS₂ can effectively protect MoS₂ from degeneration and improve the device’s performance.

4. Conclusions

In summary, MoO₂ nanoribbons and nanoflakes were grown following the APCVD method and characterized via OM, AFM, Raman spectroscopy, SEM, and TEM. The morphology of MoO₂ was modulated from a nanoribbon to a nanoflake by increasing the growth temperature. Such MoO₂ samples exhibited a standing-up growth mode instead of a lying-down one, and both their basal planes were MoO₂(100). The FET device based on Au/Ti/MoO₂ electrodes showed better electrical performance due to the protection of MoS₂ from the deposited metal electrodes. Our findings provide a convenient method for synthesizing MoO₂ and modulating its morphology. The target MoO₂ may have great potential to be used as an electrode for novel 2D material-based optoelectronic devices.