Abstract
A novel approach to convert a 2D WSe2 film into a WO x film by the laser-induced oxidation process using a nickel (Ni) adsorption layer, namely Ni-assisted laser oxidation process, was demonstrated. A layer of Ni metal was deposited to absorb the heat of the laser and oxidize the top WSe2 layer, for which an aluminum oxide (Al2O3) layer is then deposited as a barrier layer. A continuous wave laser with a wavelength of 808 nm is selected as the laser source, as it is not absorbed by the WSe2 layer. By introducing a patterned Ni layer, the selective oxidation process on WSe2 into WO3 can be achieved owing to the photothermal effects caused by the Ni layer with O2 gas. The successful oxidation parameters, including laser powers and irradiation durations with a fixed Al2O3 barrier layer thickness of 50 nm, were investigated. Resistive random-access memory (RRAM) devices fabricated using the Ni-assisted laser-oxidized WO3 structure exhibit clear resistive switching behavior compared to the structure without the laser oxidation process. In addition, the Ni-assisted laser-oxidized WO3 structure showed obvious surface enhanced Raman spectroscopy (SERS) signals of crystal violet (CV) and methylene blue (MB) with the lowest concentration of 10–6 M.
Keywords: laser oxidized WO3 , patternable laser process, WSe2 and WO3 based RRAM, SERS measurements


Introduction
Two-dimensional (2D) transition metal dichalcogenides (TMDCs) have attracted extensive interest due to their unique electrical, optical, and catalytic properties. − Unlike graphene, which lacks a bandgap, TMDCs such as MoS2, WS2, and WSe2 exhibit thickness-dependent semiconducting properties with bandgaps ranging from 1 to 2 eV, making them suitable for field-effect transistors (FETs), photodetectors, and memory devices. − Their atomic thickness, flexibility, and strong light–matter interaction enable efficient scaling and integration in van der Waals heterostructures. Among various TMDCs, tungsten diselenide (WSe2) has emerged as a remarkably versatile candidate due to its ambipolar carrier transport, high mobility, and stability under ambient conditions. − These features make WSe2 attractive for both digital and analog applications, including complementary logic circuits and neuromorphic architectures. − Moreover, the methods to modulate its properties through chemical doping, phase engineering, and localized oxidation have opened up new routes for tailoring in-plane heterojunctions and functional interfaces. − Despite these advantages, the scalable and spatially controllable patterning of 2D WSe2 layered films into different heterostructures remains a significant challenge. In particular, converting 2D WSe2 layered films into tungsten trioxide (WO3) films with high precision and compatibility with existing device fabrication processes are still under development.
The precisely controlled oxidation of 2D WSe2 layed film has proven to be an effective strategy to modulate its electrical and chemical properties, enabling applications ranging from logic transistors to neuromorphic devices. − For instance, self-limiting thermal oxidation has been shown to selectively convert an atomically thin 2D WSe2 layer into an ultrathin WO3 film while preserving lateral resolution, facilitating in-plane heterojunction engineering. , Likewise, oxidized WSe2-based FETs have demonstrated improved performance because of well-defined interfaces between semiconducting and insulating regions. Among all oxidation techniques, a conventional laser-assisted oxidation process provides a rapid and mask-free route to locally transform WSe2 into WO3 using a focused laser beam irradiated under an ambient atmosphere. However, the conventional laser oxidation often suffers from limited spatial control due to thermal diffusion, resulting in poorly defined oxide boundaries. Additionally, the lack of site-selective activation limits its utility in the precise patterning process. Recent efforts have explored lateral WSe2/WO3 heterostructures for memristive and synaptic applications, underscoring the promise of such systems for neuromorphic electronics. Nevertheless, achieving deterministic and high-resolution heterointerfaces remains a technical barrier.
To address these limitations, we present a systematic investigation on the Ni-assisted laser oxidation process to convert TMDs into metal oxide systems, where 2D WSe2 films were chosen for the idea demonstration. By introducing a patterned Ni layer, the oxidation process can be selectively achieved at designated sites due to a photothermal effect caused by the Ni layer. This approach offers improved spatial selectivity and controllable oxidation compared to conventional laser oxidation, since the Ni layer acts as a localized heat absorber that confines the reaction region and minimizes unwanted damage to adjacent areas. Structural and compositional results using Raman spectroscopy, atomic force microscopy (AFM), transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS) confirm the successful and localized formation of the WO3 films from the 2D WSe2 layered films by the Ni-assisted laser oxidation process. Then, the effects of laser parameters and oxygen atmosphere for the Ni-assisted laser oxidation process were investigated. Meanwhile, unlike previous studies that relied on blanket oxidation or uncontrolled spot irradiation, our method enables the formation of complex and well-resolved oxide patterns, such as the NTHU pattern, with clear chemical contrast. Moreover, RRAM devices based on the Ni-assisted laser-oxidized WO3 structure exhibit pronounced resistive switching behavior compared to those without laser oxidation. The observed on/off ratio is about 100 at an operating voltage of 1 V for the Ni/Al2O3/laser oxidized-WO3/W structure, while no switching behavior was observed for the Ni/Al2O3/WSe2/W structure without the Ni-assisted laser oxidation process and the Ni/Al2O3/WO3/W structure that the WO3 film was only deposited by E-beam evaporation. In addition, the same structure demonstrates clear SERS responses, showing distinct signals for both crystal violet (CV) and methylene blue (MB). This Ni-assisted laser oxidation strategy offers a scalable, mask-free method for spatially programmable oxide patterning in 2D materials. Our work provides a new direction for engineering 2D heterostructures through Ni-assisted, site-selective laser processing, with potential applications in memory devices and integrated molecular sensing platforms.
Results and Discussion
Figure illustrates the initial characterization of the WO3 film formed by the Ni-assisted laser oxidation process from 2D WSe2 layered films. This method highlights the role of the Ni adsorption layer in facilitating oxygen diffusion and reaction at the surface of the 2D WSe2 layered film. The detailed fabrication processes of the Ni (50 nm)/Al2O3 (50 nm)/WSe2 (40 nm) stacking and the laser oxidation process are shown in Figure S1a. Note that to avoid the reaction of Ni and WSe2 layered film, a Al2O3 film prepared by the E-gun deposition was used as the barrier layer, and the uniform WO x film was deposited by E-gun after the deposition of the Al2O3 film. Figure S1b,c display an optical microscopy (OM) image and a Raman spectrum of the uniform WO x film. Then, the 2D WSe2 layered films were synthesized by the plasma-assisted selenization process from the E-gun deposited WO x film, for which the heating and cooling curves of the plasma-assisted selenization process are shown in Figure S1d. After that, the samples were loaded into a laser vacuum chamber, which can apply oxygen gas inside, to carry out the laser oxidation process. The continuous wave (CW) laser with a wavelength of 808 nm was used for the oxidation process. For the selection of metal as the laser absorption and heating layer, Ni layer has been demonstrated to exhibit the highest absorption behavior at a wavelength of 808 nm compared with other metals by the Beer–Lambert law. The detailed parameters for sample preparation and the laser oxidation process are described in the method section. To confirm that most of the laser will be absorbed by the Ni layer, we also compared the light absorption of Al2O3 and WSe2 films, as shown in Figure b. Ni still shows the highest absorbance above the Al2O3 and WSe2 layers at the wavelength of 808 nm, which indicates that the Ni layer will trigger the oxidation reaction on the WSe2 film with O2 gas to form a WO3 film by a thermal annealing process because of a photothermal effect on the Ni layer. Note that the function of the Al2O3 layer acts as a separation layer between the WSe2 and Ni layers. Therefore, the overall laser oxidation process is dominated by the underlying Ni layer, which is called the Ni-assisted laser oxidation process. The OM images and corresponding Raman spectra before and after Ni-assisted laser oxidation are provided in Figure c, revealing the successful oxidation of WSe2 into WO x . After the Ni-assisted laser oxidation process, there is an obvious color change with a circular pattern in the upper OM image, suggesting that the oxidation areas, namely the WO3 area. The Raman spectra confirm that the phase should be WO3, with which peaks at 701 and 801 cm–1 were measured at the laser-irradiated area, while Raman signals of the only WSe2 peak at ∼250 cm–1 appeared before the Ni-assisted laser oxidation process, corresponding to the OM image without any color difference. To confirm that the surface of the WSe2 film remained intact after the Ni-assisted laser oxidation process, an atomic force microscope (AFM) was used to compare the surface roughness before and after the Ni-assisted laser oxidation process, as shown in Figure d. The values of R a before and after the Ni-assisted laser oxidation process are 0.53 and 0.54 nm, respectively, and the corresponding AFM images also show no significant difference on surface morphology, suggesting that the surface is not destroyed by the Ni-assisted laser oxidation process.
1.
(a) Schematic diagram of WSe2 oxidation by the Ni-assisted laser oxidation process. (b) Absorption spectra of Ni, Al2O3, and WSe2 films. (c) An Optical image and corresponding Raman spectra of the WSe2 film before and after the Ni-assisted laser oxidation process. (d) AFM images of the WSe2 film before and after the Ni-assisted laser oxidation process.
To gain deeper insight into the spatially resolved effects of the Ni-assisted laser oxidation process on the WSe2 film, detailed characterizations were performed. An OM image and corresponding Raman spectra from three distinct regions, including inside the laser-irradiated area (red spot), at the boundary of the laser-irradiated area (blue spot), and outside the laser-irradiated area (gray spot), are shown in Figure a. From the Raman spectra, strong WO3 Raman signals emerge inside the laser-irradiated area and the boundary, while pristine WSe2 peaks dominate the region outside the laser irradiation zone. Raman mapping images of the WO3 characteristic peak at 800 cm–1 at the center of the laser-irradiated area with an area of 100 × 100 μm2 are presented in Figure b, which shows uniform oxidation results within the laser-irradiated area. The uniform oxidation can be expected due to the uniform thermal heating provided by the underlying Ni layer, which significantly reduces the effect of the nonuniform radical distribution of the laser beam during the laser irradiation process. Furthermore, we investigate detailed parameters of the Ni-assisted laser oxidation, including laser power and laser irradiation durations (time), while fixing the thicknesses of WSe2, Al2O3, and Ni layers of 40, 50, and 50 nm under the O2 gas of 50 sccm, as plotted in Figure c. Note that OM and Raman results were utilized to determine all the results. The × patterns represent conditions without any reaction after the Ni-assisted laser oxidation process, revealing that the laser power is too low or the irradiation time is insufficient to trigger the oxidation reaction from WSe2 to WO3 film. The purple triangular patterns predominantly appear at laser parameters with high power and a damaged surface after applying a laser power of 23 W for 60 s, as shown in Figure S2a. The corresponding Raman spectra after applying the laser power of 23 W with the irradiation time of 60 s also show no material left at the center of the damaged surface, although the WSe2 could be successfully oxidized at the undamaged place (Figure S2b). Most importantly, the red star patterns represent suitable laser parameters that can successfully oxidize WSe2, as indicated by the Raman and OM results after the Ni-assisted laser oxidation process. Furthermore, it can be applied to the line-scan process from our CW laser system. The schematic diagram and corresponding parameter table are shown in Figure S2c, for which the Y-scale represents the scan rate of the laser and the X-scale represents the power of the laser. Based on the table, the laser power of 19 W with laser scan rates of 0.5 and 1 mm/s can achieve the oxidation of the WSe2 into the WO3, indicating that a higher laser power is needed for the line scan oxidation process than the point scan process. Apart from the laser scanning mode, we investigated the effect of the different Al2O3 thicknesses between the Ni and WSe2 layers. As the thickness of the Al2O3 oxide layer increases to 100 nm, longer irradiation time and higher laser power are needed to successfully oxidize the WSe2 into WO3, as shown in Figure S2d. The reason can be explained by the lower heat transfer efficiency from the Ni layer to WSe2.
2.
(a) Optical image and corresponding Raman spectra of the positions inside the laser spot, at the boundary of the laser spot, and outside the laser spot. (b) Raman mapping images of the WO3 peak at the laser annealing position. (c) The laser oxidation results were acquired under various laser powers and irradiation times, using a laser point. (d) Cross-section TEM images of positions inside the laser spot (red spot), at the boundary (blue spot), and outside the laser spot (gray spot).
Furthermore, the transmission electron microscopy (TEM) analysis was used to investigate the WSe2 and WO3 structures after the Ni-assisted laser oxidation process. The cross-sectional TEM analyses were carried out at three representative positions in the OM image of Figure a, including inside the laser spot (red spot), at the boundary (blue spot), and outside the irradiated area (gray spot), respectively (Figure d). In the center region, a distinct WO3 layer with a thickness of ∼30 nm with a crystalline structure was observed (Figure d1). The visible lattice spacing values are about 0.37 nm, which highly correspond to the crystallized structure of WO3. The boundary region at the blue spot site showed partial oxidation, with a mixed-phase structure of WO3 and residual WSe2. The WO3 phase appeared at the left-hand side, and the WSe2 phase is mainly observed on the right-hand side, according to the location of the blue spot in Figure a and d2. Moreover, the TEM image outside the irradiated area shows the crystalline structure of WSe2 with a lattice spacing value of 0.68 nm for the entire layer (Figure d3). Elemental analysis by TEM-energy dispersive spectroscopy (EDS) line scanning (Figure S3) further supported these structural transitions. The line scan profiles show no remaining Se after the Ni-assisted laser oxidation process, while the atomic ratio of W and O is about 30% and 70%, respectively.
Furthermore, X-ray photoelectron spectroscopy (XPS) was utilized to analyze the composition of the samples before and after the Ni-assisted laser oxidation process. In Figure a, the XPS spectrum of W at the surface of the WSe2 film after the Ni-assisted laser oxidation process shows W6+ valence states of W 4f spectra with binding energies located at 37.3 and 35.1 eV. Figure b shows XPS depth profiles, with which the atomic ratios of W, O, and Se were calculated from Figure c,d. The atomic ratios of W and O, being 30% and 70%, were confirmed, which are consistent with the results of TEM-EDS line scan profiles (Figure S3). The sputtering rate was 24 nm/min measured with a Si/SiO2 standard sample. The evolution of the W 4f spectra after the Ni-assisted laser oxidation process (Figure c) shows that the W 4f5/2 (37.3 eV) and W 4f7/2 (35.1 eV) peaks of WO3 appear from surface to depth of 60 nm calculated from the sputtering rate while the evolution of the corresponding Se 3d spectra (Figure d) shows the disappearance of Se in whole layer after the Ni-assisted laser oxidation process. Note that the peaks of the W4+ valence state represent the remaining bonds of W and Se, providing additional evidence of the complete oxidation of WSe2 into WO3. As shown in Figure S4a, the XPS O 1s spectra of the WSe2 film before and after the Ni-assisted laser oxidation process clearly reveals the emergence of oxygen-related signals. Before laser irradiation, no detectable oxygen peak is observed, indicating the absence of oxidation. After the laser oxidation process, a distinct O 1s peak appears at 529.4 eV, corresponding to the lattice oxygen of WO3. In contrast, Figure S4b presents the O 1s signal from the underlying Al2O3 layer, located at 530.8 eV. The clear distinction between these two peaks confirms the successful formation of WO3 from WSe2 after the Ni-assisted laser oxidation process. Furthermore, the binding energies of W 4f and Se 3d spectra for the WSe2 film before the Ni-assisted laser oxidation process were measured, as shown in Figure S5. The W 4f5/2 and W 4f7/2 peaks corresponding to the W4+ state are located at 33.4 and 31.2 eV, respectively. Meanwhile, the Se 3d2/3 and Se 3d2/5 peaks associated with Se2– appear at binding energies of 54.1 and 53.3 eV. These signals remain consistent from the surface to a deeper depth of 60 nm, confirming the complete WSe2 film before the Ni-assisted oxidation process. The correlative results from TEM, TEM-EDS line scan, and XPS affirm that the Ni-assisted laser oxidation process enables spatially selective and chemically distinct oxidation of WSe2 into WO3.
3.
(a) Binding energy of W 4f on the surface of the laser spot. (b) XPS atomic ratio depth profile after laser oxidation and binding energy spectra of (c) W 4f. (d) Se 3d.
Furthermore, the impact of oxygen flux during the Ni-assisted oxidation process was investigated. The OM image shows a high contrast of a circular pattern formed after the Ni-assisted oxidation process, which is similar to the results, as shown in Figure a, and the corresponding Raman spectra show the peak of WO3 formed by the Ni-assisted oxidation process with the higher oxygen flux than the previous experiment (Figure a,b). The XPS and TEM analyses were utilized to investigate the structure of the WSe2 film after the Ni-assisted oxidation process under the higher oxygen flow rate (100 sccm). The binding energy spectra of W 4f verify the compositions of each W valence state, with the depth of the film increased by applying the Ar+ etching process. The Se 3d peak indicates the remaining amount of WSe2 in the whole structure (Figure c,d). The evolution of W 4f spectra shows higher W4+ peaks at 30.8 and 33 eV, and Se peaks at 53.1–54 eV, with a depth of 60 nm (Figure d). Interestingly, despite the increase in the supply of O2, the thickness of the WO3 oxide after the Ni-assisted oxidation process was observed to be thinner to that formed under the lower O2 flow rate with identical Ni-assisted oxidation process (60 s and 15 W). The rapid formation of the WO3 layer at the surface of the WSe2 film under a high oxygen flow rate likely suppresses further oxidation by blocking diffusion of oxygen into the underlying WSe2 film. Such self-limiting oxidation behavior results in a thinner WO3 film after the Ni-assisted oxidation process. The cross-sectional TEM image and the corresponding EDS line scanning results, as shown in Figure e,f, reveal that the resulting WO3 layer is thinner compared to that formed under the low O2 flow rate condition. This observation highlights a key kinetic difference between the oxidation process under low O2 flow rate and O2-rich environments. While low O2 flow rate allows more sustained oxygen diffusion into the film, high-flow conditions promote rapid surface passivation. These findings imply that the oxide layer thickness is not only governed by oxygen availability, but also by the interfacial kinetics and structural integrity of the initially formed WO3. Thus, tuning the oxygen flow rate offers a feasible pathway to systematically control the oxide film morphology and thickness.
4.
(a) Optical image of the Ni/Al2O3/WSe2 film applied with a CW laser under a high O2 flow rate. (b) A Raman spectrum of WO3 formed by the laser oxidation under a high O2 flow rate. XPS depth profile of WO3/WSe2 film formed by laser oxidation under high O2 flow rate. Binding energy spectra of (c) W 4f and (d) O 1s. (e) Cross-section TEM image of the film formed by laser oxidation under high O2 flow rate. (f) TEM-EDS line scan profiles on the laser oxidized WO3 structure.
Since the formation of the WO3 film can be precisely controlled by the underlying Ni layer after the Ni-assisted oxidation process, it allows us to precisely form different patterns of the WO3 layers, depending on how we create the patterns of the underlying Ni layer by lithography methods, including photolithography or e-beam lithography. It means that the dimension of the laser beam will not limit the spatial resolution of the WO3 layer, while it will be limited by the resolution of the underlying Ni layer. To demonstrate this concept, patterned Ni layers were prepared by photolithography to form an identical patterned WO3 layer from the WSe2 film after the Ni-assisted oxidation process. Here, Raman mapping images were conducted across the boundary between the underlying Ni-deposited and the non-Ni-deposited regions, followed by the Ni-assisted laser oxidation process. An OM image (Figure a) shows the color contrast, indicating the formation of WO3 oxidized from the WSe2 film on the Ni-deposited regions after the Ni-assisted laser oxidation process. At the same time, there is no color change on the WSe2 area without the underlying Ni-deposited regions after the Ni-assisted laser oxidation process. The corresponding Raman mapping image (Figure b) shows the mapping intensity of the WO3 characteristic peak (∼801 cm–1), which is significantly enhanced in the region with the underlying Ni layer deposited below Al2O3 and WSe2 films. In contrast, the areas without the Ni-deposited layer show no WO3 signals under the identical Ni-assisted laser oxidation process. The diameter of the laser spot is about 700 μm, which is larger than the size of the OM image in Figure a (20 × 20 μm2), indicating that the laser source should irradiate the whole area. This spatial contrast strongly supports that the assistance of Ni promotes localized oxidation of WSe2. To further confirm the effect of Ni heating on the laser oxidation process, we applied the highest power (27 W) of the CW laser to the samples without a Ni layer under the Al2O3/WSe2 film. The OM image shows no damage after the laser irradiation (Figure S6a). The corresponding Raman spectrum reveals that there is only a WSe2 signal at 250 cm–1, and does not show any WO3 peak in Figure S6b. The Ni layer likely acts as a laser-absorbing layer or facilitates thermal conduction during the laser irradiation, thereby enhancing oxidation kinetics in its vicinity. In addition, we prepared Ni patterns with “NTHU” by a photolithography method to prove that the Ni-patterned regions confined the formation of the WO3. The high contrast Raman mapping images of the NTHU pattern after the Ni-assisted laser oxidation process are shown in Figure c. The successful formation of a well-defined “NTHU” pattern through spatially controlled laser oxidation further highlights the high patterning resolution and selectivity of the proposed Ni-assisted method. Such capability demonstrates not only the scalability and compatibility of the process with device-relevant geometries but also its potential for localized functionality in future optoelectronic and sensing applications. The schematic diagram of the mechanism for the Ni-assisted laser oxidation process in this work is presented in Figure d1 to d3. Since the absorbance of Al2O3 and WSe2 at the laser wavelength of 808 nm is much lower than that of the Ni layer, the CW laser is mainly absorbed by the Ni layer (Figure d1). As the heat from Ni layer transfers upward to the WSe2 layer, the O2 gas in the chamber reacts with the WSe2 layer, forming WO3 (Figure d2). The transformation process can be understood in terms of a localized photothermal mechanism. Upon laser irradiation, the Ni layer efficiently converts optical energy into thermal energy due to its high absorption and moderate thermal conductivity. The generated heat is confined at the Ni layer, resulting in a steep temperature gradient across the vertical stack. This localized heating raises the interfacial temperature above the activation energy required for oxidation, while the surrounding areas remain below this threshold. Consequently, the oxidation of WSe2 to WO3 occurs only at the Ni-covered regions, providing high spatial selectivity. During the process, the Se atoms in WSe2 are replaced by oxygen supplied from the O2 gas. The reaction can be described as a thermally activated anion-exchange process, where the local temperature rise promotes the diffusion of oxygen species into the WSe2 lattice and the outward migration of selenium. This results in the formation of a WO3 layer with a thickness determined by the balance between thermal diffusion and oxidation kinetics. Moreover, the presence of the Ni layer not only defines the heat distribution but also stabilizes the interfacial temperature, minimizing lateral thermal diffusion and structural damage. As shown in Figure d3, the oxidation boundaries are confined to the Ni-patterned regions, confirming the spatially selective and energy-efficient nature of the Ni-assisted laser oxidation process. This mechanism underlies the controllability and reproducibility of the transformation, distinguishing it from conventional laser oxidation methods that often suffer from uncontrolled heat spreading. Besides the influence of the underlying Ni layer, the selection of the substrates is also another critical issue for this oxidation method. In whole experiments, quartz substrates were selected for the laser oxidation experiments due to their low thermal conductivity, optical transparency, and chemical inertness. Compared to other substrates, such as SiO2/Si substrate, quartz enables more efficient heat localization during the laser irradiation process. If the SiO2/Si substrate was used, the Si will absorb the CW laser, which is not suitable for the Ni-assisted laser oxidation process. , The OM image shows that the film gets destroyed, and the corresponding Raman spectrum shows no WO3 formation after the laser oxidation process, as shown in Figure S6c,d. The high thermal conductivity of Si limits the efficiency of laser-induced thermal oxidation by rapidly dissipating heat away from the WSe2 surface. In addition, the nontransparent nature of silicon precludes the backside laser irradiation. These factors may impose constraints on oxidation uniformity and cause failures on patterning applications.
5.
(a) OM image and (b) Raman mapping of WO3 peak position at the edge of Ni after laser oxidation. (c) Raman mapping images of the WO3 peak on the NTHU patterns. (d) Schematic diagram of the mechanism of the Ni-assisted laser oxidation process.
The ability to locally induce the formation of the WO3 film via the patterned-Ni layers enables the creation of functional domains with distinct electrical and optical properties. , To evaluate the practical applications of such selectively oxidized regions, we fabricated the Ni/Al2O3/WO3, Ni/Al2O3/WSe2, and Ni/Al2O3/laser-oxidized-WO3 stackings into RRAM structures and compared their electronic transport characteristics. As shown in Figure a–c, the Ni layer was used as the bottom electrode, and the W electrode was deposited by e-beam deposition as the top electrode for the final step of device preparation. Figure a shows the current–voltage (I–V) switching characteristics of devices, which corresponds to the Ni/Al2O3/laser oxidized-WO3/W structure in the inset of Figure a. The high resistive state (HRS) and low resistive state (LRS) of the device were observed, with a set voltage of approximately −20 V, and a reset voltage close to 20 V under the current limit at 0.01 A. The on/off ratio can be observed to be 100 at the operating voltage of 1 V for the Ni/Al2O3/laser oxidized-WO3/W structure. The set and reset changes observed in the I–V curves indicate a filament-type switching mechanism. The I–V curve in Figure b was measured from the Ni/Al2O3/WSe2/W structure, where the laser oxidation process is not applied to the WSe2 layer. There is no switching behavior observed in the structure since the Al2O3 layer after the Ni-assisted laser oxidation process does not support the formation of the filamentary switching or the charge-trapping behaviors under the applied bias conditions. The possible reason is that the thermal treatment in oxygen-rich conditions may have passivated oxygen vacancy sites within the Al2O3 layer, suppressing the formation of conductive filaments typically required for resistive switching in RRAM devices. Additionally, the interface between Al2O3 and WSe2 may lack the defect states or electrochemical activity needed to sustain the bistable resistance state. To further confirm the benefits of the Ni-assisted laser oxidized WO3 for RRAM performance, we prepared the Ni/Al2O3/WO3/W sample by e-gun deposition without any annealing process. The I–V characteristic shows no obvious set or reset phenomenon that represents RRAM behavior in Figure c. These comparative results demonstrate that neither pristine WSe2 nor e-gun-deposited WO3 layers alone exhibit significant switching behavior. Note that only the laser-oxidized WO3 structures exhibit RRAM characteristics, confirming the importance of controlled oxidation pathways and heterointerface engineering in enabling device functionality.
6.
Current–voltage (I–V) characteristics of the forming process of the devices (a) with the laser irradiation process, (b) without the laser irradiation process, and (c) the WO3 film without selenization and laser irradiation process. (d) Raman spectra of CV on the laser-oxidized WO3 sample and (e) Raman spectra of MB on the laser-oxidized WO3 sample. (f) Schematic diagram of the SERS mechanism on WO3 and WSe2 structures.[32] Raman mapping of the WO3 pattern at the peak of 808 cm–1 after SERS measurements.
Beyond its advantages on RRAM applications, the laser-oxidized WO3 structure also exhibits promising optical properties that enable its use in sensing applications. To assess the potential of the laser-oxidized WO3 structures for sensing applications, surface-enhanced Raman scattering (SERS) measurements were conducted using crystal violet (CV) and methylene blue (MB) as sensing molecules. As shown in Figure d, the Ni-assisted laser-oxidized WO3 exhibits strong and well-defined Raman peaks corresponding to CV molecules with a concentration of 10–6 M, in contrast to the WSe2 without applying the Ni-assisted laser oxidation process, which does not show Raman signals of CV molecules. Moreover, Figure e presents a similar result using MB molecules (10–6 M) as the Raman sensing target, revealing that the Ni-assisted laser-oxidized WO3 structure exhibits strong and well-defined Raman peaks of MB molecules, further validating the generality of the enhancement effect. It should be noted that the excitation laser power is a critical parameter in determining the SERS enhancement performance. At low excitation powers, the local electromagnetic field is insufficient to produce a strong Raman signal, resulting in poor signal-to-noise ratios and higher detection limits. Increasing the excitation power leads to a proportional increase in Raman intensity due to stronger plasmon-molecule coupling and enhanced photoinduced charge transfer between the analyte and the WO3 domains. The consistent signal amplification observed across different molecules suggests that the SERS enhancement is not molecule-specific but rather originates from the intrinsic electronic and surface properties of the Ni-assisted laser-oxidized WO3. Figure f presents a schematic illustration of the proposed SERS enhancement mechanism between the Ni-assisted laser-oxidized WO3 and WSe2 without the Ni-assisted laser oxidation structures. The observed SERS enhancement can be attributed primarily to a charge transfer mechanism, given the dielectric nature of WO3. Specifically, the energy levels of WO3 align favorably with the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of the probe molecules such as crystal violet (CV) and methylene blue (MB). This band alignment facilitates photoinduced electron transfer between the WO3 substrate and the adsorbed molecules under laser excitation, thereby increasing the polarizability of the molecules and enhancing their Raman scattering cross-section. Notably, the Raman mapping image of the WO3 peak at 801 cm–1 in Figure g was obtained after CV exposure and SERS measurements, yet the characteristic oxide peaks remain visible and spatially confined. This result indicates that the laser-induced WO3 regions are chemically stable and structurally robust, withstanding molecular adsorption. The ability to spatially pattern SERS-active domains via the Ni-assisted laser oxidation process not only demonstrates structural control but also aligns with current efforts in developing multiplexed and addressable sensing platforms. Patterned SERS substrates have emerged as a powerful tool for location-specific molecular detection, enabling on-chip and CMOS-compatible applications in biosensing and environmental monitoring. These results show that Ni-assisted laser oxidation not only enables spatially defined patterning of oxide domains but also imparts them with functional electrical and optical properties.
Conclusion
In summary, we have demonstrated a Ni-assisted laser oxidation approach for spatially controlled transformation of WSe2 film into WO3 film by the high optical absorbance of Ni at 808 nm, the localized heating during the laser irradiation facilitates efficient oxidation of WSe2 under controlled oxygen atmospheres. Structural characterizations such as OM, Raman, AFM, TEM, and XPS confirm the formation of crystalline WO3 within the irradiated regions. Besides, the Ni-assisted laser oxidation process exhibits spatial selectivity, as evidenced by patterned oxidation confined to Ni-deposited areas. Furthermore, the oxidation behavior is strongly influenced by oxygen flow rate. Under high oxygen conditions, a self-limiting oxidation phenomenon occurs, where a rapidly formed WO3 surface passivates further reaction, leading to thinner oxide layers. This observation shows the critical role of interfacial kinetics and initial oxide structure in determining final film morphology. Functionally, the Ni-assisted laser-oxidized WO3 structures exhibit enhanced electrical performance with resistive switching characteristics, demonstrating their viability for RRAM applications. In addition, the laser oxidized regions show significant SERS activity toward molecules such as CV and MB, enabling ultrasensitive molecular detection. The SERS enhancement can be attributed to charge-transfer mechanisms facilitated by oxygen vacancies and the hybrid interface between WO3 and WSe2. Overall, this work provides a scalable and versatile platform for engineering oxide structures through metal-assisted laser oxidation, offering new possibilities in device fabrication, photonic sensing, and multifunctional material integration.
Experimental Section
Synthesis of Ni/Al2O3/WSe2 Stacking Films
Ni (50 nm), Al2O3 (50 nm), and WO3 (40 nm) layers were deposited on a quartz substrate by e-beam evaporation in order. Before e-beam evaporation, the quartz substrate was cleaned in a sequence of acetone, isopropyl alcohol, and deionized water. After the deposition of a Ni layer, the sample was placed at the middle stage in the chamber of the vertical selenization furnace, and selenium (Se) granules were placed in the top container of the selenization furnace. After introducing the sample and Se granules, the vertical tube was first pumped under a pressure of 9 × 10–3 Torr. Then, a mixed carrier gas of N2 (50 sccm) and H2 (100 sccm) was introduced from the top side of the furnace to carry the Se gas downward. The furnace was heated to 450 °C and maintained at the same temperature during the synthesis process with a fixed gas flow. When the temperature reached 450 °C, the plasma was opened at the same time and kept at a power of 150 W during the synthesis process. After the process, the H2 gas and plasma were turned off, and the chamber was naturally cooled down to 80 °C. Then, the chamber was vented to cool down to room temperature, allowing the sample to be removed from the stage.
Laser Oxidation Process
The WO3 structure was synthesized by a laser system with a continuous wavelength of 808 nm. The sample was transferred to a chamber, and the pressure of the chamber was pumped to below 1 × 10–2 Torr. O2 gas was introduced to the chamber when it reached the set pressure. After the pressure stabilized, the laser was applied to the sample, reacting with the O2 gas to convert WSe2 into WO3.
Material Characterization
Raman spectra and mapping images of the WSe2, WO3 film, and CV, MB molecules for SERS were acquired on the Andor Kymera 328i spectrograph with a 532 nm wavelength excitation laser, and the beam size of the laser is 2 μm with a 100× objective lens. The surface morphology of samples before and after laser irradiation was analyzed by AFM (Bruker, Dimension Icon). The chemical compositions, chemical bonding, and electronic structures were determined using XPS (PHI 5000 Versaprobe II), and the XPS spectra were calibrated by the binding energy of the C 1s peak at 284.5 eV. HRTEM analyses were carried out to examine the atomic structure of WSe2, WO3, WSe2, and WO3 hybrid layers by TEM with the Cs-corrector (JEOL JEM-ARM200F).
Device Fabrication and Electrical Measurements
A thickness of 80 nm tungsten (W) was deposited by e-beam evaporator on the WSe2, e-gun deposited WO3, and laser-oxidized WO3 film as the top electrode. The bottom Ni layer served not only as the heating layer but also as the bottom electrode for the RRAM device. Moreover, the Al2O3 layer was served as the switching layer. The switching characteristics were measured by an Agilent B1500A Semiconductor Device Parameter Analyzer in a probe station.
Supplementary Material
Acknowledgments
The research is supported by National Science and Technology Council of Taiwan through grants No. 114-2628-E-007-001-, 114-2923-E-007-005-MY3, 114-2823-8-007-005-, 114-2218-E-007-022-MBK, 114-2119-M-007-015-MBK, and 113-2112-M-007-034-MY3. Y.-L.C. greatly appreciates the use of the facility at the Instrument Center at National Tsing Hua University and National Synchrotron Radiation Research Center (NSRRC).
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.5c17500.
Process flow of WSe2 fabrication and formation of WO3 film by the Ni-assisted laser oxidation process; optical image and corresponding Raman spectra of the Ni/Al2O3/WSe2 film applied with the highest laser power; the laser oxidation results obtained under different laser power and laser irradiation times with laser line-scan; the laser oxidation results acquired under different laser power and laser irradiation times with 100 nm Al2O3; TEM-EDS line scan profile on laser oxidized WO3 structure; binding energy of W 4f and Se 3d before laser illumination; optical image and Raman spectra of the WSe2 film without Ni absorption layer after laser illumination; optical image and Raman spectra of the WSe2 film with laser irradiation applied on Si/SiO2 substrate (PDF)
Y.C. Hsu and Y.L. Chueh conceived and coordinated the study. Y.C. Hsu, R.H. Cyu, Y.Q. Huang, C.T. Chen, and P.C. Lai performed the experiments and data analysis. Y.C. Hsu constructed the figures. Y.L. Chueh provided theoretical guidance. All authors discussed the results and commented on the manuscript. Y.C. Hsu and Y.L. Chueh wrote the manuscript with contributions from all the coauthors.
The authors declare no competing financial interest.
References
- Manzeli S., Ovchinnikov D., Pasquier D., Yazyev O. V., Kis A.. 2D transition metal dichalcogenides. Nat. Rev. Mater. 2017;2(8):17033. doi: 10.1038/natrevmats.2017.33. [DOI] [Google Scholar]
- Chhowalla M., Shin H. S., Eda G., Li L.-J., Loh K. P., Zhang H.. The chemistry of two-dimensional layered transition metal dichalcogenide nanosheets. Nature Chem. 2013;5(4):263–275. doi: 10.1038/nchem.1589. [DOI] [PubMed] [Google Scholar]
- Gu X., Yang R.. Phonon transport in single-layer transition metal dichalcogenides: A first-principles study. Appl. Phys. Lett. 2014;105(13):131903. doi: 10.1063/1.4896685. [DOI] [Google Scholar]
- Wang Q. H., Kalantar-Zadeh K., Kis A., Coleman J. N., Strano M. S.. Electronics and optoelectronics of two-dimensional transition metal dichalcogenides. Nat. Nanotechnol. 2012;7(11):699–712. doi: 10.1038/nnano.2012.193. [DOI] [PubMed] [Google Scholar]
- Jariwala D., Sangwan V. K., Lauhon L. J., Marks T. J., Hersam M. C.. Emerging device applications for semiconducting two-dimensional transition metal dichalcogenides. ACS Nano. 2014;8(2):1102–1120. doi: 10.1021/nn500064s. [DOI] [PubMed] [Google Scholar]
- Butler S. Z., Hollen S. M., Cao L., Cui Y., Gupta J. A., Gutiérrez H. R., Heinz T. F., Hong S. S., Huang J., Ismach A. F.. et al. Progress, challenges, and opportunities in two-dimensional materials beyond graphene. ACS Nano. 2013;7(4):2898–2926. doi: 10.1021/nn400280c. [DOI] [PubMed] [Google Scholar]
- Wang L., Xu D., Jiang L., Gao J., Tang Z., Xu Y., Chen X., Zhang H.. Transition metal dichalcogenides for sensing and oncotherapy: status, challenges, and perspective. Adv. Funct. Mater. 2021;31(5):2004408. doi: 10.1002/adfm.202004408. [DOI] [Google Scholar]
- Huang J.-K., Pu J., Hsu C.-L., Chiu M.-H., Juang Z.-Y., Chang Y.-H., Chang W.-H., Iwasa Y., Takenobu T., Li L.-J.. Large-area synthesis of highly crystalline WSe2 monolayers and device applications. ACS Nano. 2014;8(1):923–930. doi: 10.1021/nn405719x. [DOI] [PubMed] [Google Scholar]
- Li H., Wu J., Yin Z., Zhang H.. Preparation and applications of mechanically exfoliated single-layer and multilayer MoS2 and WSe2 nanosheets. Acc. Chem. Res. 2014;47(4):1067–1075. doi: 10.1021/ar4002312. [DOI] [PubMed] [Google Scholar]
- Zhou H., Wang C., Shaw J. C., Cheng R., Chen Y., Huang X., Liu Y., Weiss N. O., Lin Z., Huang Y.. et al. Large area growth and electrical properties of p-type WSe2 atomic layers. Nano Lett. 2015;15(1):709–713. doi: 10.1021/nl504256y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Allain A., Kis A.. Electron and hole mobilities in single-layer WSe2 . ACS Nano. 2014;8(7):7180–7185. doi: 10.1021/nn5021538. [DOI] [PubMed] [Google Scholar]
- Cheng S., Zhong L., Yin J., Duan H., Xie Q., Luo W., Jie W.. Controllable digital and analog resistive switching behavior of 2D layered WSe2 nanosheets for neuromorphic computing. Nanoscale. 2023;15(10):4801–4808. doi: 10.1039/D2NR06580K. [DOI] [PubMed] [Google Scholar]
- Yu L., Zubair A., Santos E. J., Zhang X., Lin Y., Zhang Y., Palacios T.. High-performance WSe2 complementary metal oxide semiconductor technology and integrated circuits. Nano Lett. 2015;15(8):4928–4934. doi: 10.1021/acs.nanolett.5b00668. [DOI] [PubMed] [Google Scholar]
- Lee Y. T., Jeon P. J., Han J. H., Ahn J., Lee H. S., Lim J. Y., Choi W. K., Song J. D., Park M. C., Im S.. et al. Mixed-dimensional 1D ZnO–2D WSe2 van der Waals heterojunction device for photosensors. Adv. Funct. Mater. 2017;27(47):1703822. doi: 10.1002/adfm.201703822. [DOI] [Google Scholar]
- Ma Y., Liu B., Zhang A., Chen L., Fathi M., Shen C., Abbas A. N., Ge M., Mecklenburg M., Zhou C.. Reversible semiconducting-to-metallic phase transition in chemical vapor deposition grown monolayer WSe2 and applications for devices. ACS Nano. 2015;9(7):7383–7391. doi: 10.1021/acsnano.5b02399. [DOI] [PubMed] [Google Scholar]
- Chen J., Wang Q., Sheng Y., Cao G., Yang P., Shan Y., Liao F., Muhammad Z., Bao W., Hu L., Liu R., Cong C., Qiu Z. J.. High-Performance WSe2 Photodetector Based on a Laser-Induced p-n Junction. ACS Appl. Mater. Interfaces. 2019;11(46):43330–43336. doi: 10.1021/acsami.9b13948. [DOI] [PubMed] [Google Scholar]
- Kang W.-M., Lee S., Cho I.-T., Park T. H., Shin H., Hwang C. S., Lee C., Park B.-G., Lee J.-H.. Multi-layer WSe2 field effect transistor with improved carrier-injection contact by using oxygen plasma treatment. Solid-State Electron. 2018;140:2–7. doi: 10.1016/j.sse.2017.10.008. [DOI] [Google Scholar]
- Yamamoto M., Nakaharai S., Ueno K., Tsukagoshi K.. Self-limiting oxides on WSe2 as controlled surface acceptors and low-resistance hole contacts. Nano Lett. 2016;16(4):2720–2727. doi: 10.1021/acs.nanolett.6b00390. [DOI] [PubMed] [Google Scholar]
- Pei C., Li X., Fan H., Wang J., You H., Yang P., Wei C., Wang S., Shen X., Li H.. Morphological and Spectroscopic Characterizations of Monolayer and Few-Layer MoS2 and WSe2 Nanosheets under Oxygen Plasma Treatment with Different Excitation Power: Implications for Modulating Electronic Properties. ACS Appl. Nano Mater. 2020;3(5):4218–4230. doi: 10.1021/acsanm.0c00406. [DOI] [Google Scholar]
- Yamamoto M., Dutta S., Aikawa S., Nakaharai S., Wakabayashi K., Fuhrer M. S., Ueno K., Tsukagoshi K.. Self-limiting layer-by-layer oxidation of atomically thin WSe2 . Nano Lett. 2015;15(3):2067–2073. doi: 10.1021/nl5049753. [DOI] [PubMed] [Google Scholar]
- Yang S., Lee G., Kim J.. Selective p-doping of 2D WSe2 via UV/ozone treatments and its application in field-effect transistors. ACS Appl. Mater. Interfaces. 2021;13(1):955–961. doi: 10.1021/acsami.0c19712. [DOI] [PubMed] [Google Scholar]
- Tan C., Liu Y., Chou H., Kim J.-S., Wu D., Akinwande D., Lai K.. Laser-assisted oxidation of multi-layer tungsten diselenide nanosheets. Appl. Phys. Lett. 2016;108(8):083112. doi: 10.1063/1.4942802. [DOI] [Google Scholar]
- He H. K., Yang R., Huang H. M., Yang F. F., Wu Y. Z., Shaibo J., Guo X.. Multi-gate memristive synapses realized with the lateral heterostructure of 2D WSe2 and WO3 . Nanoscale. 2020;12(1):380–387. doi: 10.1039/C9NR07941F. [DOI] [PubMed] [Google Scholar]
- Medina H., Li J.-G., Su T.-Y., Lan Y.-W., Lee S.-H., Chen C.-W., Chen Y.-Z., Manikandan A., Tsai S.-H., Navabi A.. et al. Wafer-scale growth of WSe2 monolayers toward phase-engineered hybrid WOx/WSe2 films with sub-ppb NOx gas sensing by a low-temperature plasma-assisted selenization process. Chem. Mater. 2017;29(4):1587–1598. doi: 10.1021/acs.chemmater.6b04467. [DOI] [Google Scholar]
- Medina H., Huang C. C., Lin H. C., Huang Y. H., Chen Y. Z., Yen W. C., Chueh Y. L.. Ultrafast Graphene Growth on Insulators via Metal-Catalyzed Crystallization by a Laser Irradiation Process: From Laser Selection, Thickness Control to Direct Patterned Graphene Utilizing Controlled Layer Segregation Process. Small. 2015;11(25):3017–3027. doi: 10.1002/smll.201403463. [DOI] [PubMed] [Google Scholar]
- Thi M. P., Velasco G.. Raman study of WO3 thin films. Solid State Ionics. 1984;14(3):217–220. doi: 10.1016/0167-2738(84)90101-2. [DOI] [Google Scholar]
- Ramana C., Utsunomiya S., Ewing R., Julien C., Becker U.. Structural stability and phase transitions in WO3 thin films. J. Phys. Chem. B. 2006;110(21):10430–10435. doi: 10.1021/jp056664i. [DOI] [PubMed] [Google Scholar]
- Londoño-Calderon A., Dhall R., Ophus C., Schneider M., Wang Y., Dervishi E., Kang H. S., Lee C.-H., Yoo J., Pettes M. T.. Visualizing grain statistics in MOCVD WSe2 through four-dimensional scanning transmission electron microscopy. Nano Lett. 2022;22(6):2578–2585. doi: 10.1021/acs.nanolett.1c04315. [DOI] [PubMed] [Google Scholar]
- Thomas, S. M. ; Jennings, M. R. ; Sharma, Y. ; Fisher, C. ; Mawby, P. . The impact of oxygen flow rate on the oxide thickness and interface trap density in 4H-SiC MOS capacitors. In Materials Science Forum; Trans Tech Publ, 2015. [Google Scholar]
- Said-Bacar Z., Leroy Y., Antoni F., Slaoui A., Fogarassy E.. Modeling of CW laser diode irradiation of amorphous silicon films. Appl. Surf. Sci. 2011;257(12):5127–5131. doi: 10.1016/j.apsusc.2010.11.025. [DOI] [Google Scholar]
- Ahmmed, M. S. ; Talei, T. ; Hawkes, E. R. ; Song, L. ; Wenham, S. ; Wales, S. . Thermal simulation of laser annealing for hydrogenation of c-Si in solar cells. In Proceedings of the 52nd Annual Conference of the Australian Solar Energy Society, 2014. [Google Scholar]
- Lv Q., Tan J., Wang Z., Gu P., Liu H., Yu L., Wei Y., Gan L., Liu B., Li J.. et al. Ultrafast charge transfer in mixed-dimensional WO3‑x nanowire/WSe2 heterostructures for attomolar-level molecular sensing. Nat. Commun. 2023;14(1):2717. doi: 10.1038/s41467-023-38198-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Park S., Cho K., Jung J., Kim S.. Annealing effect of Al2O3 tunnel barriers in HfO2-based ReRAM devices on nonlinear resistive switching characteristics. J. Nanosci. Nanotechnol. 2015;15(10):7569–7572. doi: 10.1166/jnn.2015.11138. [DOI] [PubMed] [Google Scholar]
- Moram S. S. B., Byram C., Soma V. R.. Femtosecond laser patterned silicon embedded with gold nanostars as a hybrid SERS substrate for pesticide detection. RSC Adv. 2023;13(4):2620–2630. doi: 10.1039/D2RA07859G. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.






