Enhancing Multifunctionalities of TransitionMetal Dichalcogenide Monolayers via Cation Intercalation Yifei Yu,† Guoqing Li,† Lujun Huang,† Andrew Barrette,‡ Yong-Qing Cai,§ Yiling Yu,†,‡ Kenan Gundogdu,‡ Yong-Wei Zhang,§ and Linyou Cao*,†,‡ †
Department of Materials Science and Engineering and ‡Department of Physics, North Carolina State University, Raleigh, North Carolina 27695, United States § Institute of High Performance Computing, A*STAR, Singapore 138632 S Supporting Information *
ABSTRACT: We have demonstrated that multiple functionalities of transition-metal dichalcogenide (TMDC) monolayers may be substantially improved by the intercalation of small cations (H+ or Li+) between the monolayers and underlying substrates. The functionalities include photoluminescence (PL) efficiency and catalytic activity. The improvement in PL efficiency may be up to orders of magnitude and can be mainly ascribed to two effects of the intercalated cations: p-doping to the monolayers and reducing the influence of substrates, but more studies are necessary to better understand the mechanism for the improvement in the catalytic functionality. The cation intercalation may be achieved by simply immersing substrate-supported monolayers into the solution of certain acids or salts. It is more difficult to intercalate under the monolayers interacting with substrates stronger, such as as-grown monolayers or the monolayers on 2D material substrates. This result presents a versatile strategy to simultaneously optimize multiple functionalities of TMDC monolayers. KEYWORDS: hydrogen ion, lithium ion, intercalation, molybdenum sulfide, tungsten sulfide, molybdenum selenide, tungsten selenide
M
monolayers, ideally enhancing multiple functionalities at the same time. Here we demonstrate a versatile strategy to substantially enhance multiple functionalities of TMDC monolayers, that is, intercalation of small cations H+ or Li+ between the monolayers and underlying substrates. The intercalation may be achieved by simply immersing substrate-supported monolayers into the solutions of certain acids or salts. It is evidenced by the increase in height of the monolayer after the immersion and by a strong dependence of the functionality improvement on the interaction with substrates. The functionalities that may be improved include photoluminescence (PL) efficiency and catalytic activity. We may correlate the improvement in the PL efficiency, which may be up to orders of magnitude, to two effects of the intercalated cations: doping the monolayers and reducing the influence of substrates on the monolayers’ exciton
onolayers of transition-metal dichalcogenide (TMDC) materials, such as MoS2, WS2, MoSe2, and WSe2, bear great potential to deliver versatile functionalities due to the natural integration of many interesting features, including atomically thin dimension, semiconducting nature, perfect surface passivation, and catalytic activities.1−4 For instance, the monolayers are poised to enable the development of photonic devices such as high-efficiency LEDs and low threshold lasers because of exceptionally strong exciton binding energy and perfect surface passivation.5−9 The monolayers are also promising for application in solar-driven photocatalytic water splitting, as they may efficiently absorb solar light10 and catalyze the hydrogen evolution reaction.11 However, the functionalities observed at TMDC monolayers often fall short of the expectation. Just as examples, the luminescence efficiency of the monolayers is usually observed in the range of 0.1−5%,12,13 and the catalytic activity is also not satisfactory.11,14−16 In order to explore TMDC monolayers for developing high-performance devices, it is necessary to find strategies to significantly enhance the functionalities of the © 2017 American Chemical Society
Received: July 11, 2017 Accepted: August 29, 2017 Published: August 29, 2017 9390
DOI: 10.1021/acsnano.7b04880 ACS Nano 2017, 11, 9390−9396
Article
www.acsnano.org
Article
ACS Nano
Figure 1. Substantial PL change of TMDC monolayers after TFSI treatment. The mapped PL amplitude of (a, b) as-grown monolayer WSe2 and (c, d) transferred monolayer WSe2 before and after TFSI treatment. The mapped PL amplitude of (e, f) as-grown monolayer WS2 and (g, h) transferred monolayer WS2 before and after TFSI treatment. All the monolayers are on sapphire substrates. The amplitude is plotted in logarithm scale as indicated by the color bar.
Figure 2. P-doping effect of the cations to TMDC monolayers. (a) XPS spectra of MoS2 monolayers before and after TFSI treatment. Typical PL spectra of (b) as-transferred and TFSI-treated monolayer MoS2, (c) as-transferred and TFSI-treated monolayer WS2, (d) as-transferred and TFSI-treated monolayer MoSe2, and (e) as-transferred and TFSI-treated monolayer WSe2. (f) Raman spectra of MoS2 monolayers before and after TFSI treatment. The dashed lines indicate the Raman peak positions of the MoS2 monolayers before TFSI treatment.
treatment. The PL exhibits obvious change after the treatment, but the tendency is different for the WSe2 and WS2 monolayers. The PL of the as-grown WSe2 monolayer increases by more than 1 order of magnitude (Figure 1a,b), while the transferred WSe2 monolayer shows obvious PL decrease after the treatment (Figure 1c,d). In contrast, both the as-grown and transferred WS2 monolayers show similarly substantial PL enhancement after the treatment (Figure 1e,h). We have also studied MoS2 and MoSe2 monolayers and find both exhibiting changes similar to WS2 monolayers with the PL improved by orders of magnitude after the TFSI treatment (Figure S1). Additionally, sulfuric acid and Li-TFSI may enable similar change in the PL of TMDC monolayers as TFSI (Figure S2). We may correlate the treatment-induced PL change to the cations H+ or Li+. This is supported by the negligible change in PL at the monolayers treated with the solvent (DI water used as the solvent for sulfuric acid; 1,2-dichlorobenzene and 1,2dichloroethane as the solvent for TFSI or Li-TFSI) (Figure S3) or with the solutions of Na-TFSI, K-TFSI, or Na2SO4, which have the same anions but different cations compared to TFSI, Li-TFSI, or sulfuric acid (Figure S4). Note that previous studies have also reported significant PL enhancement at WS2 and MoS2 monolayers after TFSI treatment,20−22 but the results of
dynamics, while more studies are necessary to better understand the mechanism for the improvement in the catalytic functionality.
RESULTS AND DISCUSSION We start with illustrating the dramatic improvement in PL efficiency caused by the cation intercalation. The intercalation of cations was achieved by immersing the monolayers into solutions of certain acids or lithium salts. We first examine the PL of the TMDC monolayers treated with acids (sulfuric acid or TFSI) or related lithium salt (Li-TFSI). Two types of monolayers are studied: as-grown monolayers and monolayers transferred onto another substrate after the growth. The monolayers were grown using well-established chemical vapor deposition (CVD) processes17,18 and the transfer followed a surface-energy-assisted transfer process we previously developed.19 We have previously confirmed no degradation in quality at the monolayers during the transfer.19 Sapphires were used as the growth and transfer substrates for the convenience of comparison. The monolayers were immersed into solutions of the acids or lithium salt for minutes and blow dried prior to characterizations. Figure 1a−h shows the PL mapping collected from WS2 and WSe2 monolayers before and after TFSI 9391
DOI: 10.1021/acsnano.7b04880 ACS Nano 2017, 11, 9390−9396
Article
ACS Nano
as-grown monolayer WSe2 shows improvement in PL intensity and blueshift in PL peak after the treatment (Figure 1a,b), rather than decrease and redshift as expected from the p-doping effect. Second, while the majority part of the as-grown WS2 monolayers show blueshift after the treatment as expected, some areas of the as-grown monolayer show negligible blueshift (which suggests mild p-doping effect) but still very strong improvement in PL intensity (Figure S7). These conflicts indicate that the cations must generate other effects than the pdoping for the as-grown monolayers. To find out the other effects, we explore a better understanding for the effects of the cations H+ and Li+. Very importantly, our experimental results indicate that the treatment-induced PL change is mainly contributed by the cations intercalated between the monolayer and underlying substrates, while the cations adsorbed on top of the monolayer may only have minor contribution as illustrated in Figure 3a.
our work bear some difference from the previous studies. First of all, we can exclude any observable change in the density of sulfur vacancies during the treatment, as XPS measurements show negligible change in the stoichiometric ratio (Figure 2a and Figure S5). This is different from the result of the previous studies, which shows a decrease in the density of sulfur vacancies of MoS2 after the treatment.20−22 Additionally, we observe PL enhancement at WSe2 (Figure 1b) and MoSe2 monolayers (Figure S1) after the treatment, which is different from the result of the previous studies showing no PL enhancement at selenide monolayers.20−22 We also observe similar PL enhancements at the monolayers treated by TFSI, LI-TFSI, or sulfuric acid (Figure S2), while previous studies mainly focus on TFSI.20−22 Our experimental results indicate that one major reason for the treatment-induced PL change lies in the p-doping to the monolayers by the cations H+ or Li+. The p-doping to the monolayer is evidenced by PL, XPS, and Raman measurements (Figure 2a−d) and also supported by first principle simulations (Figure S6). While the XPS measurement indicates no change in the stoichiometric ratio after the treatment, the peaks show an obvious shift toward lower energies. For instance, the S 2p and Mo 3d peaks of the TFSI-treated monolayer MoS2 shift around −0.12 eV (Figure 2a), consistent with what was previously observed at p-doped MoS2.23−25 The p-doping is also supported by the shift of PL. Generally, the PL of the monolayers MoS2, MoSe2, and WS2 blueshift (Figure 2b−d), while the monolayer WSe2 shows redshift after the treatment (Figure 2e). This shift in opposite directions can be ascribed to the different intrinsic doping of the monolayers. Monolayers MoS2, MoSe2, and WS2 are intrinsically n-doped, and p-doping may lower the density of free electrons to promote the emission of neutral excitons over negatively charged excitons (trions), which may lead to blueshift in PL as trions emit at longer wavelengths.13,26,27 In contrast, monolayer WSe2 is intrinsically p-doped, and p-doping may increase the density of holes to further promote the emission of positively charged excitons (trions), giving rise to PL redshift.13 Raman measurements support the p-doping as well. For instance, the A1g Raman peak of monolayer MoS2 blueshifts by ∼0.5 cm−1, but the E12g peak shows no visible change at all after the treatment (Figure 2f). This is because that the A1g peak is subject to a stronger influence of electron−phonon coupling than the E12g peak and tends to stiffen (blueshift) with p-doping.28 Additionally, our first principle calculations confirm that H+ and Li+ may indeed p-dope TMDC monolayers by attracting electrons from the monolayers (Figure S6). However, the p-doping is not the only effect that the cations H+ and Li+ may generate to affect the PL. The p-doping effect may well explain the treatment-induced PL change observed at the transferred monolayers (Figure 2b−e), but it alone cannot explain the PL change observed at the as-grown monolayers. As discussed in the preceding text, the p-doping effect may promote the emission of neutral excitons over trions for the monolayer that are intrinsically n-doped, including MoS2, MoSe2, and WS2, but suppresses the emission of neutral excitons for the intrinsically p-doped monolayer WSe2. As the emission efficiency of trions is much lower than that of neutral excitons,13 the PL is expected to increase at monolayers MoS2, MoSe2, and WS2 but decrease at monolayer WSe2 after the treatment. This nicely matches what was observed at the transferred monolayers (Figures 2b−e), but shows some conflict with the result of the as-grown monolayers. First, the
Figure 3. Cation intercalation and dominant role of the intercalated cations. (a) Schematic illustration for the intercalation of cations and the dominant role of the intercalated cations. (b) AFM measurements for the height of a typical as-grown WS2 monolayers before (blue curve) and after (red curve) the TFSI treatment. Inset is a typical AFM image. The mapped PL of a typical monolayer on SiO2/Si substrates prepatterned with holes. The central circular area is the PL signal collected from the suspended area and the other part collected from the supported part. Inset is the optical image of a typical suspended monolayer with a scale bar of 3 μm.
The intercalation is evidenced by an increase in the height of the as-grown monolayers from around 0.70 nm to around 1.1 nm after the TFSI treatment (Figure 3b), as the presence of intercalated species may increase the separation between the monolayer and the substrate. Note that the transferred monolayers do not show an increase in height after the treatment (Figure S8), because the as-transferred monolayer, which has a height of around 1.1 nm, may have enough separation from the substrate to accommodate the intercalated species. The intercalation and the dominant role of the intercalated cations are also evidenced by the strong dependence of the PL change on the presence of substrates. We transferred TMDC monolayers onto substrates with prepatterned holes (Figure 3c inset) and then performed TFSI treatment. Remarkably, the PL of the substrate-supported area increases by orders of magnitude after the treatment, while only a very mild enhancement of no more than 50% can be observed 9392
DOI: 10.1021/acsnano.7b04880 ACS Nano 2017, 11, 9390−9396
Article
ACS Nano
Figure 4. Substrate-dependent treatment effect for TMDC monolayers. (a) Optical image of monolayer WSe2 with some part on h-BN (yellow color) and others on sapphire substrates (gray color). PL mapping collected from the sample (b) before and (c) after TFSI treatment. The dashed curve indicates the position of the h-BN substrate. (d) Optical image of a monolayer MoS2 transferred onto as-grown WS2 on SiO2/Si substrates. The shaded area indicates the as-grown WS2, and the dashed white lines represent the edge of the transferred monolayer MS2. PL mapping collected from the sample (e) before and (f) after TFSI treatment. The dashed black lines indicate the MoS2 on the WS2. (g) Optical image of a monolayer MoS2 with some part on graphene and others on SiO2/Si substrates. The shaded area indicates the MoS2 monolayer. The dashed black line and the arrow indicate the area covered by graphene. Mapped PL intensities collected from the sample (h) before and (g) after TFSI treatment.
at the PL of the suspended area (Figure 3c,d). If the cations adsorbed on top of the monolayer play an important role, then we would expect substantial improvement in PL at the suspended area as well. Additionally, the intercalation and the dominant role of the intercalated cations are further supported by the dependence of the PL change on the interaction of the monolayer with substrates. We systematically examined the treatment-induced PL change at a large number of transferred and as-grown monolayers, each >50 monolayers. The transferred and as-grown monolayers show dramatically different reproducibility and uniformity in the PL change. Only a small portion (