1 Introduction
Inspired by the significant advances of graphene, the transition metal dichalcogenides (TMDs) with finite bandgaps have begun to receive increasing attention in the last decade [
1–
3]. Unique optical and electrical properties of these TMDs, including massive excitonic effect [
4,
5], remarkable many-body interaction [
6,
7], and giant spin-valley coupling [
8,
9], make them promising candidates for next-generation optoelectronic devices. However, the intrinsic defects, such as vacancies and impurities for as-prepared monolayers, generally result in the extremely low photoluminescence (PL) quantum yield (QY, ranging from 0.01 % to 6 % at room temperature [
10]), obscuring the further development of high quantum efficiency optoelectronic devices [
11]. Recently, many research efforts have been devoted to enhance the PL intensity of monolayer WS
2, which may be distinguished into four possibilities: (i) surface passivation with organic superacids [
10,
12,
13], (ii) energy transfer from quantum dots (QDs) [
11,
14–
16], (iii) plasmonic coupling through metallic nanostructures [
17], and (iv) lifting emission rate by photonic crystals [
18,
19].
Surface passivation is a widely known method that can tune the optical properties of two-dimensional (2D) materials. To date, the nonoxidizing organic superacid bis (trifluoromethane) sulfonamide (TFSI) [
10,
20], oleic acid (OA) [
13], and sodium sulphide (Na
2S) [
12] have been used to treat as-grown WS
2 monolayer. The total PL has been enhanced by over two orders of magnitude, yielding near-unity PL QY in monolayer TMDs. On the other hand, perovskite (CsPbBr
3 [
16], BA
2MA
3Pb
4I
13 [
15]), semiconductor (InGaN) [
11], and heterostructural core/shell (WO
3) [
14] QDs have been adopted to improve the PL intensity of WS
2. The mechanism is attributed to the energy transfer from QDs to TMDs, or the aggregation of excitons in WS
2. Other studies have sought to boost PL intensity by enhancing light-matter interaction and exciton-coupled surface plasmon polaritons. Fei
et al. have demonstrated that the PL intensity of WS
2 can be improved by more than one order of magnitude in the Ag nanowire/WS
2/Ag film hybrid structure [
17]. Very recently, over 300-fold enhancement on PL of monolayer WS
2 has been determined with a tailored photonic crystal structure by Jiang’s group [
19]. This enhancement is attributed to the improvement of PL efficiency and the acceleration of spontaneous emission rate. However, the harsh nature of chemical treatment or plasmonic coupling necessitates additional fabrication steps [
13]. Furthermore, contaminants and impurities are always inevitably introduced during the fabrications, which may significantly reduce the performance of optoelectronic devices based on these 2D materials [
21]. Therefore, achieving high and stable PL intensity of monolayer WS
2, through an all-optical scheme without any chemical treatments and additional fabrication steps, is still an exploratory of fundamental research.
In this work, we report the giant and flexible enhancement of PL emission in monolayer WS2 by an all-optical scheme with femtosecond laser irradiation. More than two orders of magnitude enhancement of PL intensity as compared to the as-prepared sample had been readily determined. To reveal the underlying mechanism, we performed PL spectra under different conditions. To optimize the enhancement effect, we investigated the enhancement of PL intensity as a function of irradiation time under various power densities. We also examined the stability of this enhancement by storing the processing samples in a vacuum chamber filled with nitrogen gas. At last, we demonstrated various micropatterning on monolayer WS2 based on this giant, flexible, and robust PL enhancement. Benefiting from these features, our approach provides a new platform for the fabrication of information storage and nanophotonic devices based on these 2D materials.
2 Results and discussion
Monolayer WS
2, purchased from Six Carbon Technology Shenzhen, was synthesized on the SiO
2/Si substrate by chemical vapor deposition (CVD) [
22]. The quality of the as-prepared samples has been characterized by optical microscopy and Raman spectroscopy. Their monolayer nature with a thickness of ~0.8 nm has been confirmed by atomic force microscopy, as shown in Fig. S1 in the electronic supplementary materials (ESM). The main optical experiments, including femtosecond laser irradiation, PL imaging, and spectra measurements, were performed by using a home-built scanning confocal microscope, as described in the previous works [
23,
24], which can also be found in Fig. S2 in ESM. In particular, a femtosecond laser (Mai Tai DeepSee) at the center wavelength of 780 nm with a pulse width of about 70 fs was used to irradiate and engineer the optical properties of monolayer WS
2. To probe the PL intensity and spectra of as-prepared and post-processed WS
2, a 405 nm continuous-wave (CW) laser was used and collimated with the femtosecond laser. To avoid thermal damage or further engineering, the excitation power density of the CW laser was limited to low regimes (~0.21 kW/cm
2). Monolayer WS
2 was placed on a motorized three-dimensional piezoelectric ceramic stage. The micro-patterning was achieved by moving the WS
2 sample with respect to the focused laser beam in a programmable way.
Very recently, some groups have engineered and promoted optical properties of monolayer MoS
2 and WS
2 by CW lasers with high power density. The PL intensities were significantly enhanced by a factor of ~10-fold [
25–
27]. However, the engineered times were generally in ten minutes or even hours. To further improve PL performance and significantly shorten the engineering time, we employ the femtosecond laser with high peak power density to serve as a new platform. Figure 1 presents the typical PL image after femtosecond laser irradiation. The irradiation power density was set to 2.1 MW/cm
2, the irradiation time was set from 30 ms to 30 s, as the solid arrows shown in Fig. 1(a). PL image of each irradiated spot was enlarged in Fig. 1(b). Some of the engineered spots present asymmetric profile. Two possible reasons are responsible for this phenomenon. The inhomogeneous quality of monolayer WS
2, confirming by their inhomogeneous PL intensity (Fig. S3), is assumed to be the main reason. Another rational factor is the asymmetric distribution of the laser beam after focusing by the objective (Profiles of the focused laser beams have been presented in Fig. S4). From Fig. 1(b), we can find that PL intensities of post-engineered areas are gradually enhanced before
ms [Fig. 1(b)]. Then PL intensities of the center area sharply decrease even quench to the background. These results can be well understood. The center of the laser spot had a higher power density. Thus the crystal structure of monolayer WS
2 was primarily damaged by femtosecond laser irradiation. To quantify the enhanced result, we define the enhancement factor (
EF) as the ratio of the integrated PL intensity of the post-engineered area to that of the as-prepared sample. We calculated the enhancement factor for the area of full engineered spots [
EF1, assuming the diameter of the largest spot to be 3 μm, as the solid line shown in the inset of Fig. 1(c)] and the center area (
EF2, as the dashed line shown in the inset), the results are shown in Fig. 1(c). The maximum enhancement factor under these conditions reaches a value up to 60 times, significantly higher than the engineered results under CW laser irradiation [
21,
28].
To elucidate the underlying mechanism, we measured and plotted PL spectra of engineered spots (the center area) as a function of irradiation time, as shown in Fig. 2(a). With the prolonging of irradiation time, PL spectra were firstly enhanced and then quenched. The blueshift of the PL peak position can be clearly observed, as indicated by the dotted lines. The spectral profile of the as-prepared sample shows a broad distribution, ranging from 1.60 eV to 2.10 eV, as the top panel shown in Fig. 2(b). While after laser irradiation, the PL peak with the highest energy was dramatically enhanced and predominated the full spectrum, as shown in the bottom panels. All the spectral profiles can be well deconvoluted into three components, as the shadows shown in Fig. 2(b). According to the previous works [
27,
29,
30], we can attribute the high and media energies to the exciton (A
0) and trion (A
T) emission. The energy difference between A
0 and A
T is between 27 meV to 36 meV, reasonably coinciding with the trion dissociation energy reported in Refs. [
30,
31], as the bottom panel shown in Fig. 2(c). The low energy peak can be assigned to the defect-bound localized excitons (A
D) [
29,
30,
32], which shows significant spectral broadening due to complex grain boundary defects and nanosize defects with different sizes [
33]. The evolution of grain boundaries and nanosize defects during femtosecond laser irradiation results in the variation of its peak energy [
34], as the triangle date shown in Fig. 2(c). Interestingly, similar evolutions are also observed for full width at half maximum (FWHM) of A
D, as shown in Fig. S5. In contrast, peak energies and FWHM of A
0 and A
T show little variation.
On the other aspect, the integrated PL intensities of A
0 and A
T present dramatic change during irradiation (Fig. S5). The ignorable intensity of A
0 of the as-prepared sample was enhanced to the dominating component after laser irradiation for more than 100 ms. To quantify the intensity evolution, we plotted the spectral weight (the ratio of the integrated PL intensity of each component to the total intensity) as the function of irradiation time, as shown in Fig. 2(d). Note that the weight of A
T almost stands at a plateau. The weight of A
D sharply decreases while that of A
0 rapidly increases. We speculate on two scenarios for these results. Firstly, with the prolonging of femtosecond laser irradiation, many new defects form and subsequent gas molecules (such as O
2 and H
2O) bonding to the defects convert trions to excitons by depleting electrons, which is accompanied by a remarkable enhancement of PL intensity [
35]. No significant PL enhancement under the vacuum condition provides further evidence for this hypothesis (Fig. S6). Another possible reason for such giant enhancement is that the quantum efficiency of quasiparticles (both excitons and trions) localized at the defects will be significantly enhanced, as suggested by Nan and Kim [
35,
36]. As a consequence, PL intensities of both trions and excitons are greatly improved. However, the spectral weight of trions drops slightly during irradiation, as shown in Fig. 2(d). Combining with the hypothesis of gas adsorption on new-formed defects as well as the ignored enhancement under vacuum condition, we suggest that the conversion from trions to excitons also plays a vital role in the PL enhancement. After irradiation for 2.0 s and more, PL intensities of all components quenched sharply (Fig. S5) as the crystal structure of monolayer WS
2 was damaged.
The established mechanism inspires us that PL enhancement depends on the competition between the conversion from trions to excitons and the degree of crystal damage. To optimize the enhancement factor, we collected PL intensities as a function of irradiation times under different power densities on many monolayers (Fig. S7), the processing results are plotted in Fig. 3(a). We can find that the enhancement factor reaches up to more than 200 at the power close to 8.3 MW/cm
2 and irradiation time of 1.0 s, as presented in Fig. 3(b). In our experiment, we achieved the maximum enhancement factor to be 450 (Fig. S8). To guide the effective PL enhancement, we define the maximum enhancement factor (
EFmax) and its reaching time (
Tmax), and plot these two values as a function of the irradiation power density, as shown in Figs. 3(c) and (d), respectively. Note that
EFmax firstly ascends with the increase of the power density, and then descends after the power density larger than 8.3 MW/cm
2, indicating that PL intensity cannot be further enhanced by monotonically increasing irradiation power. Unlike
EFmax, the
Tmax monotonically shortens with the increase of irradiation power density. And, more remarkable,
Tmax is generally in sub-second. As compared to the irradiation by CW lasers [
25,
26,
28], the engineered times have been shortened by three orders of magnitude, which significantly improves their practical applications. Intriguingly, the ascent and descent of
EFmax as the function of the irradiation power density can be well fitted by the mono-exponential function,
EFmax(
P) =
EFmax(0) +A·exp[-(
P-
P0)/
τ]. Here,
P is the irradiation power density in MW/cm
2,
τ represents the ascent or descent characteristic power, determining to be 1.55 and 1.60 MW/cm
2 for the ascent and descent processes, respectively. Similarly,
Tmax can be well fitted by mono-exponential function as well. The fitting result gives the shortening characteristic power of 5.80 MW/cm
2. These values offer a quantitative description to controllably design PL intensity of monolayer WS
2 by femtosecond laser irradiation.
The precise control of PL intensity allows for direct and flexible micropatterns on monolayer WS2. As shown in Fig. 4, many desired micropatterns based on the enhancement of PL intensity have been created readily, such as English letters of “WS2” and Chinese character of “光” (meaning “light”). Note that the enhancement of PL intensity can be controlled by changing both the irradiation power density and irradiation time, here we fixed the irradiation power at 8.3 MW/cm2 and irradiation time of 1.0 s for each pixel. To explore the PL enhancement, we plot the PL intensity of as-prepared and engineered area along the dashed lines shown in Fig. 4(a). The enhancement factor can be up to 50, as presented in Fig. 4(b). Furthermore, periodic arrays of microstructures (named micro-grating) can be readily created as well, as shown in Fig. 4(d). The corresponding enhancement profiles along with the yellow rectangle and their multi-Gauss fits have been presented in Fig. 4(e) (PL image of the as-prepared and the relevant fitting results can be found in Fig. S9 and Table S2). Almost uniform FWHM of engineered lines indicates the precision control of PL enhancement by femtosecond laser irradiation. This flexible micropatterning enables promising applications in optical data storage, display technology, and micro-optoelectronic devices. To further explore the promise of this technique, we characterized the stability and reliability of micropatterning after irradiation. The processing sample was stored in an improvised vacuum chamber filled with nitrogen gas. All the optical experiments were performed at the ambient atmosphere without any further protection. Figure S10 presents the corresponding PL image and PL intensity profiles along the marked lines at different time points after laser irradiation. The evolution of relevant PL enhancement factors is illustrated in Figure 4F. From which we can find that the enhancements are reasonably stable in three years. The small variation possibly originates from the change of the parameters of the home-built experimental setup (such as the excitation power and focusing of the objective) and the unavoidable degradation of monolayer WS2 (such as pollution by the dust and water vapor). This result further explores the promising applications of our scheme in the information storage and related optical devices.
3 Conclusion
In conclusion, we have shown that the PL intensity of monolayer WS2 can be enhanced by a factor of more than two orders of magnitude by femtosecond laser irradiation under optimized parameters, as compared to the as-prepared sample. The maximum enhancement is 450 times in our experiment, and the irradiation time is shortened by three orders of magnitude, as compared with the irradiation by CW lasers. This giant enhancement is attributed to not only the improvement of QY of exciton and trion when they are localized to the laser-induced defects but also the conversion from trion to exciton. In particular, this PL enhancement is robust, where the enhanced PL intensity can be stability preserved for more than three years. This giant, flexible, and robust PL enhancement allows us to create micropatterning on monolayer TMDs, which has promising applications in information storage and nano-optoelectronic devices. The enhanced PL QY also enables the development of new high-performance light-related devices, such as light-emitting diodes, lasers, and solar cells.