Nanoscale Insights into the Hydrogenation Process of Layered α‑MoO3
Weiguang Xie,† Mingze Su,† Zebo Zheng,‡ Yu Wang,† Li Gong,§ Fangyan Xie,§ Weihong Zhang,§ Zhi Luo,⊥ Jianyi Luo,∥ Pengyi Liu,† Ningsheng Xu,‡ Shaozhi Deng,‡ Huanjun Chen,*,‡ and Jian Chen*,§ †
Siyuan Laboratory, Guangzhou Key Laboratory of Vacuum Coating Technologies and New Energy Materials, Department of Physics and ⊥Department of Electronic Engineering, Jinan University, Guangzhou 510632, China ‡ State Key Lab of Optoelectronic Materials and Technologies, Guangdong Province Key Laboratory of Display Material and Technology, School of Electronics and Information Technology and §Instrumental Analysis & Research Center, School of Physics and Engineering, Sun Yat-sen University, Guangzhou 510275, China ∥ School of Applied Physics and Materials, Wuyi University, Jiangmen 529020, China S Supporting Information *
ABSTRACT: The hydrogenation process of the layered αMoO3 crystal was investigated on a nanoscale. At low hydrogen concentration, the hydrogenation can lead to formation of HxMoO3 without breaking the MoO3 atomic flat surface. For hydrogenation with high hydrogen concentration, hydrogen atoms accumulated along the direction on the MoO3, which induced the formation of oxygen vacancy line defects. The injected hydrogen atoms acted as electron donors to increase electrical conductivity of the MoO3. Near-field optical measurements indicated that both of the HxMoO3 and oxygen vacancies were responsible for the coloration of the hydrogenated MoO3, with the latter contributing dominantly. On the other hand, diffusion of hydrogen atoms from the surface into the body of the MoO3 will encounter a surface diffusion energy barrier, which was for the first time measured to be around 80 meV. The energy barrier also sets an upper limit for the amount of hydrogen atoms that can be bound locally inside the MoO3 via hydrogenation. We believe that our findings has provided a clear picture of the hydrogenation mechanisms in layered transition-metal oxides, which will be helpful for control of their optoelectronic properties via hydrogenation. KEYWORDS: hydrogenation, MoO3, hydrogen bronzes, oxygen vacancies, transition metal oxides
A
One of the most effective post-treatments for improving the physical and chemical properties of the pristine α-MoO3 is hydrogenation.9−11,16,18,20 For example, the electronic structure of the α-MoO3 can be strongly modified to give large density states within the band gap via hydrogen treatment.10,11 Incorporation of such hydrogenated MoO3 as an interfacial layer can greatly improve the performance of the organic optoelectronic devices.11 In addition, annealing the MoO3 in hydrogen-rich ambient is suggested to create oxygen deficiencies and induce gap states with high density.11,21 These gap states can on one hand dramatically enhance the electrical conductance of MoO3 and on the other hand give rise to high responsivity of photodetection in the visible region.9 A very recent study proposed that irradiation of the MoO3 nanoflakes in aqueous environment with light allowed the electronic properties of these nanoflakes to be facilely tuned
s an important member of the transition-metal oxide family, α-MoO3 has drawn considerable attention in recent years due to its significance in gas sensing,1−3 biosensing,4 electrochromism,5,6 energy storage,7,8 and optoelectronic devices.9−11 In particular, the specific energy band structure of the α-MoO3 can benefit the hole injection/ extraction, which can therefore significantly improve the performance of photovoltaic devices by incorporating αMoO3 as a hole-selecting layer.11−15 In recent years, the twodimensional (2D) layered MoO3 began to gain attention due to the intense investigation in 2D materials.16−19 For example, 2D α-MoO3 with a high dielectric constant offers a solution for high carrier mobility in thin film electronic devices by strongly suppressing the Coulomb scattering of carriers.9 Despite its scientific and technological importance, the luminescence, electronic, and optoelectronic properties of pristine α-MoO3 are, however, very poor due to its large and indirect band gap (≥3 eV).9 Therefore, it is not suitable for direct utilizations unless specific post-treatments have been applied. © XXXX American Chemical Society
Received: November 24, 2015 Accepted: December 21, 2015
A
DOI: 10.1021/acsnano.5b07420 ACS Nano XXXX, XXX, XXX−XXX
Article
www.acsnano.org
Article
ACS Nano
atoms on surface Ot site to the adjacent Ot, Oa, or Os site is around 0.6 eV. The DEB between the two layers within the double layer is smaller than 0.35 eV.25 Another calculation showed that for the diffusion between two adjacent double layers the DEB was around 0.1 eV, while it was even smaller for diffusion within the same monolayer.24 The only experimental results so far suggest that the DEB for hydrogen diffusion in HxMoO3 powder is around 0.15−0.3 eV depending on the value x (x = 0.2−2.0).27 To precisely measure the hydrogen DEB, one needs techniques capable of recording the electronic structure changes of the MoO3 during the hydrogenation process. In addition, such techniques should be of high spatial resolution to eliminate ensemble and inhomogeneous averaging. Defect formation and its influence on the optical and optoelectronic properties of the MoO3 is another issue up for debate. The hydrogenation will usually lead to the formation of metallic HxMoO3, where the injected hydrogen atoms are expected to aggregate at the bridging oxygen atoms to form the −OH2 groups at high hydrogen concentrations. The −OH2 groups are easy to desorb as H2O molecules and leave oxygen vacancies,26 giving rise to gap states that are believed to be responsible for the modifications of the sample’s optoelectronic properties.9,11 Previous studies using atomic force microscopy (AFM) and electron diffraction have observed protruding structures extending preferentially along the direction on the surface of the high temperature-hydrogenated αMoO3.28,29 These protrusions were ascribed to the formation of HxMoO3. On the contrary, another study suggested that oxygen vacancy defects instead of HxMoO3 precipitates formed along [101] direction during the hydrogenation.30,31 To the best of our knowledge, none of these studies have paid attention to how these structural changes affect the electrical and optoelectronic properties of the MoO3. To clarify these issues, one needs to directly characterize the structural changes at the nanoscale and relate them to the electrical and optoelectronic performances of the hydrogenated MoO3. Herein, we employed 2D layered α-MoO3 crystal as a model platform to study the hydrogenation process and related mechanisms. The atomic flat surface of the 2D MoO3 can favor the unambiguous real-time and in situ surface characterizations, whereby the hydrogenation dynamics and hydrogen DEB can be determined. Furthermore, combination of the nanoscale morphology and electrical and optical characterization techniques allows us to observe the hydrogenation-induced oxygen defects and establish their correlation with the electrical and optical properties of the 2D α-MoO3 crystal. On the basis of these findings, we finally discussed the local control of the MoO3 hydrogenation. We believe that the results obtained in our study have clarified the hydrogenation mechanisms of the MoO3, which can provide guidelines for controlling the electrical and optoelectronic properties of the transition-metal oxides.
due to the intercalation of the hydrogen ions. These twodimensional substoichiometric MoO3 nanoflakes can be utilized to fabricate field effect transistors with outstanding carrier mobility and a high on/off ratio.22 Moreover, intercalation of hydrogen into the layered α-MoO3 can also lead to strong plasmon resonances in the near-infrared region, where the plasmon wavelengths can be facilely tuned by changing the amount of hydrogen.19 Although this great progress has unambiguously demonstrated hydrogenation to be an exceptional manner for improving the electrical and optoelectronic properties of the α-MoO3, there is still a lack of consensus on the detailed mechanisms governing the hydrogenation process.23−26 So far, there are two important issues that need to be addressed. The first one is hydrogen diffusion on the surface and within the MoO3.24,25 As is known, the α-MoO3 is composed of layered planar units with double layers (Figure 1a). The double layer,
Figure 1. Structural characterization of the α-MoO3 crystal. (a) Schematic showing the crystal structure of the α-MoO3. The lattice constants are a = 3.963 Å, b = 13.856 Å, and c = 3.697 Å, respectively. (b) XRD spectrum of the α-MoO3. Inset shows the photograph (5×) of the laminar α-MoO3 crystal. (c) Highresolution TEM (HRTEM) image of a typical α-MoO3 2D nanoplate. Inset shows the corresponding SAED pattern. (d) Raman spectrum of the pristine α-MoO3 nanoplate.
which is composed of linked distorted MoO6 octahedra, has a thickness of around 1.4 nm.10 The hydrogen atoms will first be adsorbed on the topmost layer of the MoO3 at Ot sites. Thereafter, they will diffuse into the adjacent Ot, Oa, or Os sites, and finally into the bulk.25 During such a process, there exist diffusion energy barriers (DEB) for the hydrogen atom diffusing from one site to the other. To quantitatively determine the DEB is very important because according to these values one can design procedures to locally confine the hydrogen atom at specific sites. However, right now, the reported quantitative values of the DEB are controversial and remain to be clarified. For example, previous theoretical calculations have shown that the DEB for diffusion of hydrogen
RESULTS AND DISCUSSION In our study, the 2D α-MoO3 was prepared using thermal evaporation method (see the Methods for details). The obtained α-MoO3 was single crystalline and structured in a lamellar manner, with lengths around tens of millimeters and thicknesses around tens of micrometers (Figure 1b, inset). Xray diffraction (XRD) characterizations showed a series of sharp peaks at 13.7°, 26.6°, 39.8°, 53.6°, and 68.3°, which can be assigned to the (0k0) facets of the orthorhombic α-MoO3 B
DOI: 10.1021/acsnano.5b07420 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
through the α-MoO3 decreased when the sample was hydrogenated. The transmittance at 633 nm was reduced to only ∼33% of its original value after the sample was exposed to hydrogen for 30 min. For a small extent of hydrogenation, the color and transmittance intensity of the sample can recover after about 10 h exposure in air, while after the sample has undergone a long period of hydrogenation it cannot recover to the transparent and colorless state. In addition, distinct cracks can be observed on the surface of the α-MoO3 (Figure 2b). To further reveal the nanoscale origins of the color changes during the hydrogenation, in situ and real-time characterizations using scattering-type scanning near-field optical microscopy (sSNOM) were utilized. Morphology and near-field optical scattering intensity at 633 nm of the sample can be simultaneously obtained in the s-SNOM measurements, whereby the color changes of the α-MoO3 can be directly correlated to its structural evolvements. Normally, the hydrogenation process was too fast for SNOM imaging. In addition, the coated Pt nanoparticles can affect imaging of the intrinsic αMoO3 surface. In view of these we chose to characterize the hydrogenation process in the area next to the Pt-coated region. The hydrogenation will start at the Pt-coated region and then extend to the area without Pt nanoparticles (Figure S2, Supporting Information). In such a manner, measurements of intrinsic changes of the electrical and optical properties of αMoO3 surface can be ensured. On the other hand, in the experiments we found that the hydrogen flow significantly affected the stability of electrical and optical near-field signals, and we therefore performed the measurements after the hydrogen flow was shut off. At the beginning of the hydrogenation, the near-field scattering intensity at 633 nm only experienced small attenuation (Figure 3b). The decrease of the scattering intensity reflected increase of the light absorption by the MoO3. No obvious structural modifications can be observed. In contrast, as the hydrogenation progressed, the near-field scattering intensity decreased quickly and the MoO3 surface underwent continuous structural changes. Small protrusions with a height of ∼7 Å and width of tens of nanometers first appeared (Figure 3c). Such protrusions extended along a direction ∼45° relative to the [001] direction of the MoO3 surface, suggesting the growth direction of the protrusions was the [101] direction. As the hydrogen exposure time increased, the protrusions became higher, longer, and wider and then cracked (Figure 3d). Further hydrogenation leads to a third type of surface structure, where large ripples along the [001] direction were found (Figure 3e). At this stage, the near-field optical intensity decreased and stabilized at a minimum value (Figure 3b). Furthermore, the defect regions exhibited much weaker light scattering (Figure 3f), suggesting that the defects had stronger light absorption than the other regions on the MoO3 surface. This observation was consistent with the macroscopic image of the sample, where cracks showed a darker color (Figure 2b). Electrical transport property is another important issue involved during the hydrogenation of the α-MoO3. To reveal the nanoscale mechanism associated with the electrical property changes during the hydrogenation, we investigated the surface potential evolvement and correlated it with the macroscopic resistance changes of the MoO3. Knowledge of the surface potential can then help to establish the surface band structure of the MoO3. The surface potential was measured in situ using a Kelvin probe force microscope (KPFM) during the hydro-
(JCPDS: 05-0508). The surface of the α-MoO3 crystal was therefore indexed to the (010) plane. The transmission electron microscope (TEM) image and selected area electron diffraction (SAED) patterns indicated that the as-grown crystal had lattice spacings of 3.64 and 3.92 Å (Figure 1c), which were consistent with the interplane separations of (001) and (100) planes of the α-MoO3, respectively. These results clearly indicated that the MoO3 crystal grew along the [001] direction. AFM characterizations also revealed the layered structure of the αMoO3, with a double-layer thickness of ∼1.4 nm (Figure S1, Supporting Information). Raman spectrum of the sample showed distinct bands corresponded to the Ag, B1g, and B3g modes of the α-MoO3 (Figure 1d).32,33 Under catalysis action of the Pt nanoparticles sputtered onto the surface of the MoO3, the hydrogen molecules will be decomposed into hydrogen atoms and then spill over the (010) plane of the MoO3 (Figure 2a). As discussed above, the
Figure 2. Hydrogenation of the layered α-MoO3 crystal. (a) Schematic showing the adsorption and diffusion of hydrogen onto the surface of the α-MoO3. (b) Color changes of the α-MoO3 during hydrogenation. (c) Evolution of the Raman spectra of the αMoO3 during hydrogenation. The time labeled above the spectra indicated the corresponding hydrogenation duration. Each spectrum except the one of the pristine MoO3 was lifted up manually for better comparison. (d) Evolution of each Raman peaks during the hydrogenation. Each curve has been normalized to its maximum during the hydrogenation.
hydrogen atoms will be adsorbed at Ot sites and subsequently diffuse to the adjacent Ot, Oa, or Os site and finally into the bulk. During these processes, the appearance of the MoO3 crystal will undergo color changes from transparent to dark blue (Figure 2b). Furthermore, usually HxMoO3 will be formed along with the hydrogenation. The Raman spectra recorded during the hydrogenation process demonstrated that the intensity of all of the modes associated with the pristine MoO3 decreased rapidly. At the same time, a new mode at 1017.5 cm−1 emerged gradually (Figure 2c and d). This new mode can be assigned to the MoO stretching vibration mode of type I H0.3MoO3.34 The color changes of the layered α-MoO3 can also be manifested from its light transmission spectra at the visible region. As shown in Figure 3a, the transmittance intensity C
DOI: 10.1021/acsnano.5b07420 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
Figure 3. Optical and morphology characterizations of the α-MoO3 crystal during hydrogenation. (a) Transmittance spectra of the pristine and hydrogenated MoO3 crystal. (b) Evolution of the near-field scattering intensity of the MoO3 during hydrogenation. Data was obtained by averaging over a scanning area of 2 μm × 2 μm on the MoO3 surface. (c−e) Topography evolution of the MoO3 during the hydrogenation. (f) Optical near-field amplitude at 633 nm associated with the topography shown in (e).
hydrogen flow rate, the resistance of the α-MoO3 crystal decreased once the sample was exposed to the hydrogen. Meanwhile the surface potential was found to be elevated. At this stage, no surface structural change was observed. If we shut off the gas flow, the sample’s resistance remained stable for more than 10 h. As the hydrogen flow rate or exposure time was increased, the surface potential increased accordingly (Figure S3, Supporting Information). Because no structural changes were observed on the MoO3 surface, the increase of the surface potential can be ascribed to the formation of the HxMoO3. If we further increase the hydrogen flow rate to a critical value (50−70 SCCM), the sample’s resistance started to fluctuate strongly (Figure 4b). Furthermore, the resistance increased dramatically to a value even higher than the sample’s pristine resistance if the gas flow was shut off. At this stage, line defects appeared and started to develop into cracks on the surface of the MoO3 (Figure 3c). We therefore ascribe the increase of the resistance to these cracks, which can scatter the carriers severely. In addition, the surface potential of the MoO3 decreased (Figure S3, Supporting Information). More precisely, the distribution of the surface potential was nonuniform, where the defect sites along the [101] direction exhibited lower potential than their adjacent regions (Figure 4c,d). We should point out that due to the differences in the specific sample structure and location of the gas pipeline used in every measurement there existed divergence in the hydrogen flow rate where the surface potential of the sample started to decrease. Nevertheless, surface potential evolvements in all of the hydrogenation processes qualitatively agreed with each other. Another important issue we want to point out is that the surface potential obtained from the KPFM measurements can be affected by the status of the conductive tip. In our experiment, this will be caused by the illusions of the Pt-coated KPFM tip due to its interaction with the hydrogen flow.
genation. Simultaneously, the electrical resistance of the αMoO3 was recorded via a two-point-probe method (Figure 4a). According to the typical electrical transport curves of the Ptcoated α-MoO3 crystal, the resistance modifications can be classified into two stages (Figure 4b). In the first stage with low
Figure 4. In situ characterizations of the electrical properties of the α-MoO3 crystal during the hydrogenation. (a) Schematic showing the in situ electrical measurements. The surface potential changes was measured using the KPFM method. The electrical transport characterization was performed in situ using a two-point probe procedure. (b) Evolvement of the α-MoO3 resistance during the hydrogenation. The sample has a dimension of ∼20 mm × 1 mm × 0.03 mm. The hydrogen flow rates are indicated correspondingly to different stages. (c) Topography of the α-MoO3 when exposed to high hydrogen flow. The scan size was 5 μm × 5 μm. (d) Surface potential distribution of the α-MoO3 associated with the topography shown in (c). D
DOI: 10.1021/acsnano.5b07420 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
Figure 5. Mid-infrared s-SNOM characterizations of the α-MoO3 crystal. (a) Evolution of the near-field scattering intensity of the MoO3 during the hydrogenation. Each data was obtained by averaging over a scanning area of 2 μm × 2 μm on the MoO3 surface. (b) Topography of the MoO3 with ripples. (c) Optical near-field amplitude at 10.69 μm associated with the topography shown in (b).
Figure 6. Surface band structure of the α-MoO3 crystal. (a) Schematic showing the band bending on the surface of the α-MoO3 with HxMoO3. (b) Schematic showing the band bending on the surface of the α-MoO3 with oxygen defects. (c) Comparison of the current profiles across a typical defect line when the hydrogen flow was switched on and off. The topography profile across the defect line was also included. The crack region was indicated by the two green dashed lines.
Normally, the hydrogen flow will probably interact with the Ptcoated tip and thereafter affect the results from two aspects. First, the hydrogen flow can interfere the mechanical oscillation of the AFM tip. Such interferences will disturb the KPFM signals. We found that under a moderate hydrogen flow smaller than 200 SCCM the vibration of the AFM tip was unaffected if the flow rate was stable. Under such a scenario, the KPFM can give stable and correct surface potentials. Throughout our study, we carefully controlled the experimental conditions to guarantee that the surface potentials were obtained when the hydrogen flow rate was stable. The second possible influence of the hydrogen flow on the measured surface potential is the adsorption of hydrogen onto the KPFM tip. Under catalysis action of the Pt-coated tip the hydrogen molecules might be decomposed into hydrogen atoms and adsorbed onto the tip. Consequently, the work function of the tip will be modified, giving rise to illusions in the surface potential of the sample below. To elucidate such an effect, we have performed KPFM measurements on a pristine α-MoO3 thin film without any sputtered Pt nanoparticles. The pristine MoO3 thin film was inert to the hydrogen. On the other hand, the KPFM was conducted in tapping mode, where the contact duration of the tip with the MoO3 was short enough to eliminate hydrogen diffusion into the MoO3. Therefore, if there was any surface potential changes collected, they should be associated directly with the modifications of the work function of the tip. When the pristine MoO3 was exposed to a hydrogen flow of 0 to 50 SCCM, the surface potential experienced negligible changes, with only a small background offset of 2 meV in comparison to that of the sample before hydrogen exposure (Figure S4, Supporting Information). Such an offset was smaller than the resolution of the KPFM (10 meV) and can be neglected. The work function of the tip can therefore be seen as unchanged during the KPFM scanning under hydrogen exposure. In view of the above two points, the surface potentials measured in the
KPFM was the intrinsic ones from the sample surface, while not originated from the illusions of the tip due to its interaction with the hydrogen. To further reveal the origins of the surface potential changes, we used a mid-infrared s-SNOM technique to image the electron density distributions on the MoO3 surface during the hydrogenation. As shown by previous studies, the mid-infrared s-SNOM was able to detect local conductivity and dielectric material properties down to the tens of nanometers scale.35−39 For the mid-infrared imaging in our experiment, the SNOM tip was illuminated with a focused CO2 laser beam (λ = 10.69 μm). By detecting the scattered light in an interferometric manner, near-field optical amplitude and topography images can be obtained. Usually in doped semiconductors, the near-field optical intensity at mid-infrared was originated from the collective excitations of the free electrons (holes).37−39 Thus, regions with stronger near-field optical amplitudes represented their higher electron concentrations. The average mid-infrared near-field scattering intensity on the MoO3 surface increased gradually as the sample was continuously exposed to the hydrogen flow, indicating that the free charge carrier density was increased (Figure 5a). Such an observation was consistent with the aforementioned decrease of the sample resistance. In addition, the distribution of the nearfield scattering intensity on the MoO3 surface was nonuniform, which was similar to that of the surface potential discussed above. In comparison to their adjacent regions, the line defects exhibited a much weaker light scattering intensity, suggesting that the carrier densities at the defect sites were lower (Figure 5b and c). On the basis of the above results, the hydrogenation and its influence on the electrical and optical properties of the α-MoO3 can be understood as follows. During the initial hydrogenation, the hydrogen atoms will diffuse into the MoO3 and be mainly bound to the lattice corner-sharing oxygen atoms, whereby the E
DOI: 10.1021/acsnano.5b07420 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano HxMoO3 can be formed as a result of the O−H bonding. Under such a circumstance each hydrogen atom will contribute one free electron to the MoO3 lattice, giving rise to the increase of the free electron densities as well as the surface potential (Figures 4 and 5). In our KPFM measurements, the MoO3 crystal was grounded while the conductive tip was biased. The obtained surface potential was actually the potential difference between the tip and sample surface. Furthermore, by choosing such an experimental architecture the Fermi level (EF) of the MoO3 was kept constant. Therefore, an increased (reduced) surface potential collected suggested that the conduction band moved away from (toward) the energy state of the tip and toward (away from) the EF, resulting in downward (upward) bending of the surface energy band.40,41 In this regard, the increase of the surface potential suggests downward bending of the surface energy band (Figure 6a), while it does not correspond to EF pinned into the conduction band. As the hydrogen concentration was increased or the hydrogen exposure was prolonged, the oxygen atoms bridging two octahedra were bonded with two hydrogen atoms and led to the formation of −OH2 groups (Figure 6a). The accumulation of the −OH2 groups leads to deformation of the octahedra along the [101] direction, and therefore structural protrusions emerged. As the concentrations of the −OH2 group increased, local strain became strong enough to release the H2O molecules, leaving the oxygen vacancy line defects and resulting in the formation of MoO3−x (Figure 6b). These observations are consistent with previous calculation results.21 Release of the oxygen atoms will create oxygen vacancies. Each oxygen vacancy acts as positively charged defect, which can effectively trap the free electrons.42−44 As a result, when the hydrogenation was shut off, the free electrons were trapped around the oxygen defects, giving rise to reduction of free carrier densities and upward energy band bending (Figure 6b). These processes can further be corroborated by measuring the localized surface current across a defect line in real time during the hydrogenation. As shown in Figure 6c, when the sample was purged with hydrogen flow, a distinct increase of local current on the protrusions along [101] direction next to the defect cracks was recorded. Furthermore, the current was much higher than the regions surrounding the protrusions. This contrast was due to the accumulations of the hydrogen atoms along the [101] direction. On the other hand, the current recorded from the crack regions was minimum, which was due to the release of H2O molecules. For the color appearance, the visible s-SNOM results indicated that both of the HxMoO3 and oxygen defect lines were responsible for the coloration of the MoO3 during the hydrogenation. However, the visible light absorption due to the oxygen defects was much stronger than the adjacent areas where the HxMoO3 located. Therefore, we owe the oxygen defects to the dominant origin of the chromism of the α-MoO3 crystal. With the knowledge of the hydrogenation mechanisms, we can manage to quantitatively determine the DEB for hydrogen diffusion within the MoO3. To do so, we initiated the hydrogenation of the α-MoO3 using a Pt-coated AFM tip (Figure 7, inset). Under hydrogen flow, the AFM tip was brought into contact with the MoO3 surface. The hydrogen molecules would then be dissociated into two hydrogen atoms and diffuse from the tip into the MoO3. The hydrogenation can be controlled by tailoring the hydrogen flow rate and contact duration of the AFM tip to the sample surface. In addition, the
Figure 7. Surface potential changes by localized hydrogenation on a bulk α-MoO3 using a Pt-coated AFM tip. The inset shows the much higher center surface potential due to local hydrogenation by a Ptcoated AFM tip.
hydrogenation process could be stopped whenever the tip was lifted from the sample surface. In such a manner, the changes of the surface potential as time elapsed could be monitored with the same AFM tip, whereby the hydrogen DEB could be studied and measured. There are two types of diffusion barriers governing the hydrogenation process, which are the surface diffusion (S-DEB) and bulk diffusion barriers (B-DEB), respectively. The adsorbed hydrogen atoms will first be trapped at specific Ot sites on the surface by the S-DEB. Every trapped hydrogen atom can contribute one additional electron to the MoO3. Therefore, the surface potential was lifted due to the accumulation of the hydrogen atoms. However, once the surface potential at the contact point was large enough for the hydrogen atoms to overcome the S-DEB, the additional hydrogen atoms could diffuse to adjacent oxygen sites (Oa, or Os site) and thereafter into the body of the α-MoO3 crystal. Diffusion of the hydrogen atoms within the crystal body is limited by the B-DEB. These processes can be manifested by the changes of the surface potential on the MoO3 2D nanoplate (Figure S5, Supporting Information). Under the hydrogenation, one can observe that, besides the contact area, the surface potential of the surrounding area was also elevated. This finding indicated that when the hydrogen atoms overcame the S-DEB they could easily diffuse into the nanoplate from the contact point. Due to the finite thickness of the nanoplate (∼50 nm), the hydrogen atoms accumulated at regions underneath the sample surface, giving rise to the lifted surface potential of the MoO3 as a whole. Because all of these processes occurred at room temperature, one can conclude that the B-DEB for hydrogen atom within the internal parts of the MoO3 was small (
