1. Introduction

Two-dimensional (2D) transition metal dichalcogenide (TMDC) semiconductors have attracted significant research interest, from the viewpoint of fundamental physics and promising applications, especially in the fabrication of optoelectronic and nanophotonic devices [1–3]. However, a common roadblock in the employment of TMDC for practical applications is the limited absorption, due to their ultrathin thicknesses and confined spectral selectivity determined by their bandgaps [4–6]. A typical solution for increasing the absorption of incident light is to couple these layered materials with other nanostructures with strong absorption intensity [7,8]. Among various alternative nanomaterials, zero-dimensional (0D) colloidal semiconductors quantum dots (QD) are particularly interesting, due to their large absorption covering the broad spectrum from ultraviolet to near-infrared light [9,10]. The stringent requirement for high-performance optoelectronic devices has inspired worldwide efforts with the rapid development of this kind of hybrid structure. To date, many 2D-QD hybrid structures with the superiority in the optical performance have been demonstrated by combing monolayer and few-layer 2D materials with PbS, CdSe/ZnS, ZnCdSe/ZnS, and perovskite CsPbX₃ QD [11–18]. For example, the hybrid 2D-QD phototransistor consisting of few-layer n-type MoS₂ and p-type PbS QD has been designed, with the responsivity up to 6×10⁵ A/W in the spectral region of 400 to 1500 nm, beyond the bandgap of MoS₂ [12]. On the other hand, the responsivity of another hybrid photodetector, formed by thin MoS₂ film and ZnCdSe/ZnS QD, is 3 orders of magnitude larger than the 2D photodetector itself [4].

To further improve the optical performance of 2D- and/or QD-based devices, the interface interactions in the 2D-QD hybrid structures have also been investigated. Typically, there are two possible interactions between 2D materials and QD undergoing photoexcitation: the photoinduced electron transfer and/or nonradiative energy transfer [19,20]. In these hybrid structures, QD generally acts as the donor, and 2D materials serve as the acceptor components. In this case, the photoluminescence (PL) intensity of QD is always quenched. The results demonstrated that both interactions are thickness-(for 2D materials), size-(for QD), and distance-dependent [15–17]. In particular, the rates for both electron transfer and energy transfer raise with the increase of the 2D thickness, and thus PL of QD is further weakened [19,20]. In these previous works, the thicknesses of both 2D materials and QD films were limited to several to tens of nanometers. Under the circumstances, some interesting questions arise if one further increases the thicknesses of 2D materials and/or the QD films, and takes into account further interface effects beyond electron/energy transfer: Can we suppress the quenching of QD or even enhance their PL? If so, we can also improve the optical performance of QD in these hybrid structures. This result is particularly important for QD-based devices, due to that QD is generally thick in practical applications.

In this work, we comprehensively investigate the optical interference effect in the hybrid 2D-QD structures with different thicknesses and excitation conditions by theoretical modeling and experimental study of their PL signal. In the theoretical model, we take into account the effects of energy transfer and multiple reflections. The results demonstrate that with thin 2D materials or the QD films, the energy transfer predominates the interaction, resulting in the dramatic PL quenching of the QD films. While with the increase of the thickness, PL quenching can be suppressed and even be reversed to enhance. It’s also found that this PL enhancement can be modulated by changing the thicknesses of 2D materials and/or QD films. We uncover that this abnormal PL enhancement and modulation chiefly originates from the big difference in the refractive indexes between 2D materials and the QD films. To prove our theoretical predictions, we further investigate the PL intensities of different 2D-QD hybrid structures as functions of 2D thickness. The experimental results are in good agreement with theoretical simulations. Our work provides a simple, yet efficient method to enhance and modulate the PL intensity of QD in the hybrid 2D-QD structures for potential utilization in the QD-based light emission and solar harnessing applications.

2. Results and discussion

2.1 Theoretical model for the optical interference in the hybrid 2D-QD structures

For thin QD films and/or few-layer 2D materials, two possible interactions, nonradiative energy transfer (or Förster resonance energy transfer, FRET) and electron transfer, will predominate the optical performance of the hybrid 2D-QD structures. Normally, the energy transfer rate decays as $d_{2D - QD}^{ - 4}$ ($d_{2D - QD}^{ - 4}$ denotes the distance between 2D materials and the QD films), [21–24] while the electron transfer rate reduces exponentially with the distance [19,25,26]. In consequence, both energy transfer and electron transfer will die out rapidly with the increase of the distance (typically completely disappear within several to tens of nanometers). However, it’s difficult to make accurate predictions about the rates and thus their influence on the PL intensity of QD (see Section S1 in Supplement 1 for the formulas and discussions). For simplicity, here we only consider the influence of energy transfer on the PL intensity of the QD films. By considering 2D materials as a whole acceptor (rather than layer by layer [19]), the energy transfer efficiency can be simplified as ${\eta _{FRET}} = 1/(1\textrm{ + }{({d_{QD}} - x)^4} \cdot {R^{ - 4}})$, [23] where d_QD is the thickness of the QD films, x is the depth of the emission sources, as shown in Fig. 1(a). R is the Förster distance at which the energy transfer efficiency is 0.5. In this case, the reduced quantum yield (QY_FRET) of QD can be expressed as:

 

Fig. 1. (a) Schematic of electron transfer (ET) and Förster resonance energy transfer (FRET) between quantum dots (QD) and two-dimensional (2D) materials. E_ex and E_em represent the excitation laser and the photoluminescence (PL) emission, respectively. Blue dashed circles indicate QD with the height of x, the total thickness of the QD film is d_QD. The solid circle and ball indicate the photo-excited electrons and holes, respectively. (b) and (c) are schematics of the hybrid air/QD/2D/SiO₂/Si (abbreviated to A/QD/2D/SiO₂/Si) and glass/QD/2D/air (abbreviated to G/QD/2D/A) structures. The multiple reflections and the corresponding optical paths of E_ex and E_em are illustrated by the arrows. (d) Mapping quantum yields (QY_FRET) as functions of d_QD and the Förster distance R by equation (Eq.) 1. The inset presents calculated QY_FRET as a function of d_QD at the logarithmic scale. The Förster distance (R) is set to 5 nm, as the dashed line highlighted in the map. (e) Mapping PL enhancement factors (EF) as a function of the thickness of the QD film (d_QD) and 2D materials (d_WSe2) in the hybrid A/QD/2D/SiO₂/Si structure. The corresponding calculated parameters are listed in Table S1 and Fig. S8. (f) Calculated EF as a function of the 2D thicknesses (line cuts along the dashed lines in e and i). (g) Calculated EF as a function of the QD thicknesses (line cuts along the dashed lines in e and i). (h) Mapping EF as a function of the wavelength of the excitation laser (λ_ex) and the emission light (λ_em). (i) Mapping EF as a function of the thickness of d_QD and d_WSe2 in the hybrid G/QD/2D/A structure. (j) Mapping EF as a function of the thickness of QD and hexagonal boron nitride (h-BN) (d_h-BN) in the hybrid A/QD/2D/SiO₂/Si structure.

Download Full Size | PDF

Here d_m represents the minimum distance between QD and 2D materials. For bare QD, d_m can be treated as zero, while for core-shell QD, this value can be treated as the thickness of the shell.

Figure 1(d) presents the calculated QY_FRET as functions of d_QD and R. From the theoretical calculations, we can find that within the very short distance (such as bare QD and/or monolayer 2D materials), PL of QD is almost fully quenched, consistent with the previous reports [19,20,24]. Note that, with the increase of the QD film thickness, the influence of energy transfer will fade away. Intriguingly, although the effect of both energy transfer and electron transfer can be ignored with thick QD films and/or 2D materials, the multiple reflections and the corresponding optical interference effect will play an important role in the optical performance of the hybrid 2D-QD structures. Indeed, multiple reflections have been considered to uncover the abnormal enhancement of the Raman signal and the image contrast of graphene on the SiO₂/Si substrate [27–33]. The enhancement of the Raman signal largely originates from the big difference in the refractive indexes between graphene and SiO₂. Similarly, this big difference also occurs in the interface between the QD films and 2D materials, and thus the optical interference effect will take place in the hybrid 2D-QD structures as well. Under the circumstances, both the amplitude of the excitation laser and the emission light will be enhanced. The total intensity of PL from the QD films can be given by $I = \int_{{d_m}}^{{d_{QD}}} {{{|{{E_{ex}}(x) \cdot {E_{em}}(x)} |}^2}dx}$, where E_ex(x) and E_em(x) are the electric field amplitude of the excitation laser in the QD films with the depth of x and the amplitude of the emission light from the QD films.

Here we consider two hybrid structures: Air/QD/2D/SiO₂/Si (denoted as A/QD/2D/SiO₂/ Si), and Glass/QD/2D/Air (denoted as G/QD/2D/A), as shown in Figs. 1(b) and 1(c), respectively. To simplify the derivations, we ignore the higher-order multiple reflections (see Fig. S10 for details) and assume that the transmittance (t_ij) and reflection coefficients (r_ij) of light propagating from the medium i to the medium j follow the Fresnel law, i.e., ${t_{ij}} = 2\widetilde {{n_i}}/({\widetilde {{n_i}} + \widetilde {{n_j}}} )$ and ${r_{ij}} = ({\widetilde {{n_i}} - \widetilde {{n_j}}} )/({\widetilde {{n_i}} + \widetilde {{n_j}}} )$, respectively. $\widetilde {{n_i}}$ and $\widetilde {{n_j}}$ are the corresponding refractive indexes. Under the assumptions, the total amplitude of the excitation laser and the emission light of the QD films at the depth x in the A/QD/2D/SiO₂/Si hybrid structure can be expressed as (see Section S2 in Supplement 1 for the detailed derivations):

After considering both the energy transfer and the optical interference effect, the modulated PL intensity of QD in the hybrid 2D-QD structures can be described as:

To compare the PL enhancement and modulation before and after considering the multiple reflections, we also derive the total PL intensity of the QD films in the A/QD/SiO₂/Si hybrid structure without considering the energy transfer and the multiple reflections, I^Non (see Eq. S40 for the derived formula), and define the PL enhancement factor (EF) as EF = I/I^Non. The demonstrations of I, I^Non, and EF have been illustrated in Fig. S9. Figure 1(e) presents the map of calculated EF as the function of d_QD and 2D thickness (taking WSe₂ as an example, denoted as d_WSe2), the relevant parameters are listed in Table S1 and Fig. S8. The thickness of the layered materials in this work is generally in the region of a few nanometers to tens of nanometers, the optical properties of the thick layered nanometers are close to the bulk materials, however, for the sake of simplicity, we name them all as 2D materials. Apparently, the PL of the QD films can be well modulated by varying the thickness of both the QD films and 2D materials, with the maximum EF up to about 8. Figure 1(f) plots the calculated EF as a function of WSe₂ thickness with d_QD= 100 nm. Note that a slight PL quenching occurs to thin WSe₂, primarily resulting from the energy transfer. The maximum EF can be obtained at d_WSe2≈33 nm, indicating that a good optical performance of QD-based devices can be determined by using a relatively thick layered WSe₂. With the increase of the WSe₂ thickness, EF can be almost periodically modulated, the modulation period is close to 50 nm. Note that, with the increase of the WSe₂ thickness, the Förster distance R will slightly decrease, due to the stronger absorption spectra of WSe₂ nanoflakes. However, the influence of the variation of the absorption features on the PL enhancement can be ignored (as presented in Fig. S12), given the dense QD films used in this work. From Eq. (2)-4 and the calculated results, we can also manipulate the PL intensity of the QD films by changing the thicknesses, the refractive indexes, and the excitation coefficients of 2D materials, the comparisons between different layered materials (WS₂, WSe₂, MoS₂, and MoSe₂) have been presented in Fig. S11. On the other hand, EF can also be modulated by varying the thicknesses of the QD films (d_QD), as shown in Fig. 1(g). However, the maximum EF rapidly decays with the increase of the QD thicknesses. This phenomenon primarily originates from the strong absorption of the QD films, therefore the excitation laser cannot transit through the QD films. In this case, the multiple reflections between the QD and 2D films can be ignored. Furthermore, EF can also be manipulated by varying the wavelengths of the excitation laser and/or the emission light (i.e., different QDs), as illustrated in Fig. 1(h). Note that the enhancement behaviors originating from the variation of the excitation wavelength and the emission wavelength are symmetrical. This can be easily understood because the optical interference is strongly correlated with the wavelength of the light. We also confirm that PL enhancement and modulation can be achieved in the hybrid G/QD/2D/A structure, as the schematic and calculated results shown in Figs. 1(c) and 1(i), respectively. Note that the maximum EF in the hybrid G/QD/2D/A structure (∼6) is slightly smaller than that in the hybrid A/QD/2D/SiO₂/Si structure (∼8), the corresponding EF profiles also present a tiny difference, as presented in Figs. 1(f) and 1(g). The reduced EF essentially originates from the similar refractive indexes between the QD films and the glass, rather than the big difference between QD and air in the hybrid A/QD/2D/SiO₂/Si structure. On the other hand, the reduced difference in the refractive indexes between 2D materials and the QD films (such as hexagonal boron nitride, h-BN, with a refractive index close to 2) will also suppress the maximum EF. As illustrated in Fig. 1(j), the maximum EF in the QD-h-BN hybrid structure is only close to 3.

2.2 PL modulation of QD in A/(CdSe/ZnS)/WSe₂/SiO₂/Si structure

To confirm the aforementioned theoretical predictions, we foremost performed PL measurements of CdSe/ZnS core-shell QD capped with WSe₂ nanoflakes, as the schematic shown in Fig. 2(a). Two main superiorities are responses for us choosing this hybrid structure. Firstly, PL spectra of CdSe/ZnS and WSe₂ can be well distinguished, as presented in Fig. 2(b), allowing for the detection of each PL property independently. Secondly, energy transfer occurs in this structure, due to the overlapping between the PL spectra of QD and the absorption spectra of WSe₂. Furthermore, the nonradiative energy transfer of this hybrid structure has been investigated recently, [17] providing the critical understanding and parameters. Particularly, the diameter of the core of CdSe/ZnS is about 8.2 nm, with the shell about 2.1 nm, as characterized by transmission electron microscopy (TEM) presented in Fig. 2(c). The WSe₂ nanoflakes with different thicknesses were firstly prepared by mechanical exfoliation from the bulk crystals and then deposited on the cleaned SiO₂/Si substrates [34,35]. The presence of different 2D materials was identified by Raman spectra. As shown in Fig. 2(d), the strong peak around 250 cm^-1 can be attributed to the two first-order modes with ${A_{1g}}$ and $E_{2g}^1$ symmetries, undoubtedly confirming the presence of WSe₂ materials [36,37]. Then, the QD films were formed by spin-coating of a CdSe/ZnS-water mixture with polyvinyl alcohol (PVA) on the prepared samples, the thickness of the QD films can be controlled by varying the concentration of the mixtures and the parameters of spin-coating parameters. The thickness of the QD films can be determined by measuring the height profiles near the edge of the scratched films (Fig. S13). As illustrated in Fig. S14, we have confirmed that the difference in thicknesses between WSe₂ nanoflakes before and after spin-coating the QD films is negligible, supporting the homogeneity of the prepared QD films. Detailed information about sample preparation can be found in Methods.

 

Fig. 2. Characterizations of the prepared hybrid samples. (a) Schematic of the hybrid A/(CdSe/ZnS)/WSe₂/SiO₂/Si structure (not to scale). The excitation laser is transited from the air by an objective, the emission light is collected by the same objective. (b) Normalized absorption (Abs.) and PL spectra of CdSe/ZnS QD and monolayer WSe₂. The solid arrow indicates the excitation laser with a wavelength of 532 nm. (c) Transmission electron microscope (TEM) image of CdSe/ZnS QD. Scale bar: 10 nm. (d) Raman spectra of thicker WSe₂. Two arrows mark the peak of ${A_{1g}}$ and $E_{2g}^1$ modes. The inset presents the phonon mode from the Si substrate for calibration. (e)-(g) Optical, atomic force microscopy (AFM), and PL image of the prepared sample. The inset in f is the height profile of the selected line. The inset in g is the histogram of the PL intensities, which were analyzed by three Gauss functions. Scale bar in e-g: 10 µm.

Download Full Size | PDF

Figure 2(e) presents the optical image of a typically prepared hybrid sample. The relatively homogeneous color contrast (cyan for QD and brown for thicker WSe₂) indicates that the surfaces of both the QD films and WSe₂ nanoflake are uniform and flat. This feature can be further identified by their atomic force microscopy (AFM), as shown in Fig. 2(f). From AFM imaging, we can further determine the height of the thicker WSe₂ nanoflake (brown in Fig. 2(e)) is about 40 nm and that of the thinner one is about 20 nm (grey in Fig. 2(e)), respectively. The thickness of the QD films was determined to be about 100 nm as well. PL imaging and PL spectra were performed using a 532 nm continuous laser (see Methods for details). Note that for thick WSe₂ nanoflakes, their PL almost disappeared and thus they cannot be well determined in this experiment. On the other aspect, by using a narrow band filter (640 ± 30 nm), PL imaging of CdSe/ZnS QD has been solely acquired, as shown in Fig. 2(g). We can find that the QD films without capped WSe₂ present almost the same PL intensity, with the value of 158 ± 18 (1σ standard derivations, as the inset shown in Fig. 2(g)) counts per second (cps). For the thinner WSe₂ nanoflakes, PL is decayed to 105 ± 15 cps. Surprisingly, PL is enhanced to 318 ± 61 cps, with the averaged enhancement factor up to twice and the maximum enhancement up to 3. Due to the large thickness of the QD film (∼100 nm), the difference in the energy transfer between the QD film and the two WSe₂ nanoflakes is negligible (as discussed in Fig. 1(d)) and thus cannot be the response to the phenomenon. Therefore, the optical interference effect originating from the multiple reflections is the most probable mechanism.

To optimize the maximum PL enhancement and verify the optical interference effect, we systematically prepared the CdSe/ZnS-WSe₂ hybrid structures with different WSe₂ nanoflakes and performed PL measurements. Figure 3(a) presents the optical images of four typically prepared samples. The spin-coating parameters for the QD films are maintained, and thus the thicknesses of the QD films have kept around 100 nm, hinting by the similar color contrast in the optical image and comparable intensity in PL imaging. To explore the effect of the multiple reflections, the thicknesses of various WSe₂ nanoflakes were depicted by AFM, as the insets shown in Fig. 3(a). Apparently, PL intensities of QD capping WSe₂ nanoflakes with the thickness of 2.5 nm and 72.6 nm are significantly quenched, with EF being 0.08. While the thickness of 27.8 nm and 105.8 nm are markedly enhanced, as shown in Fig. 3(b). EF of both conditions are up to 6 times. The thickness-dependent PL modulation can also be determined by their PL spectra, as shown in Fig. 3(c). To reveal the possible interactions, we further deconvolute all the spectra by two Lorentzian functions. The peak with higher energy is attributed to the neutral exciton, X°; while the other peak is assigned to the charged exciton or trion, X^T. The determined energies of the two peaks are 1.972(8) eV and 1.943(6) eV, their energy difference (known as the trion dissociation energy) is 0.029(4) meV. These results are consistent with previous works [38–40]. The weak and irregular variations indicate no more interactions occur during the change of the WSe₂ thicknesses. Exhilaratingly, the experimental EF is reasonably consistent with the theoretical calculations, as presented in Fig. 3(e). This result further declares the optical interference effect in the hybrid 2D-QD structures and their manipulation of the PL property of QD.

 

Fig. 3. PL intensity of CdSe/ZnS QD varies as the thickness of WSe₂ nanoflakes. (a) Optical images of typical A/(CdSe/ZnS)/WSe₂/SiO₂/Si structures with different WSe₂ nanoflakes. The insets present the height profiles of the selected lines, and the thicknesses of the WSe₂ nanoflakes are also marked. Scale bar: 10 µm. (b) PL images of the corresponding structures. The excitation conditions are the same. PL intensities of the selected lines are shown in the bottom panels of each image, the dashed lines indicate the averaged PL intensities of QD without capping WSe₂. Scale bar: 10 µm. (c) PL spectra of QD capping WSe₂ nanoflakes with different thicknesses. All the spectra are deconvoluted into two peaks (neutral exciton, X°, and charged exciton, X^T). The photon energies of X° and X^T, as well as their difference, are presented in (d). The color shadows represent the corresponding averaged values, associated with the standard derivations. (e) The comparison between experimental (Exp.) EF and the theoretical predictions (Cal.).

Download Full Size | PDF

2.3 PL modulation of QD in various hybrid structures

To further unfold the modulation of PL and corroborate the theoretical predictions, we also performed PL measurements on the other different hybrid structures. We first examine the PL modulation as a function of the QD thicknesses. As shown in Fig. S15, at a fixed thickness of WSe₂ nanoflakes, PL intensities of the QD films capping WSe₂ experienced from quenching to enhancement with the increase of the QD thickness. When the thicknesses of the QD films were larger than 400 nm, the influence of WSe₂ nanoflakes is negligible, as shown in Fig. 4(a). The tendency of PL modulation as a function of the QD thickness is also in reasonably good agreement with the calculation, as shown in Fig. 4(b). Then we replaced the WSe₂ nanoflakes with MoS₂ and performed PL measurements on the QD films, as shown in Fig. S16. PL modulation can be distinctly determined from its PL imaging (Fig. 4(c)), the corresponding EF are also in agreement with the theoretical predictions, as presented in Fig. 4(d). This result indicates that PL enhancement and modulation of the QD films in the hybrid A/QD/2D/SiO₂/Si structure can be determined by using different TMDC materials.

 

Fig. 4. PL enhancement of the QD films as functions of different parameters. (a) PL image of the QD film with a thickness of about 400 nm, the WSe₂ nanoflakes shown in the inset was buried under the QD film. The thickness of the nanoflakes was 52 nm. (b) Experimental EF as a function of the QD thickness, the dashed line is the calculated curve with d_2D= 52 nm. (c) PL image of the CdSe/ZnS/MoS₂ hybrid structure. The inset presents the optical image of the MoS₂ nanoflakes. The thicknesses and PL intensities of these nanoflakes have been illustrated in Fig. S16. (d) Experimental EF as a function of the MoS₂ thickness, the dashed line is the calculated curves with d_QD= 40 nm. Scale bars: 40 µm.

Download Full Size | PDF

To further declare that the main conclusions are applied to other hybrid structures, we also replaced CdSe/ZnS with the near-infrared CdSeTe/ZnS core-multishell QD with the emission wavelength at 800 nm. Similar PL modulation, no doubt, occurs again in this hybrid structure, as shown in Fig. 5(a). The EF values of 1.6 and 1.9 have been determined for the thickness of 15.3 nm and 29.1 nm, as presented in Fig. 5(b). The corresponding EF can be predicted by our multiple reflections model as well (see Fig. S17 for details). We then performed PL measurements on the hybrid G/QD/2D/A structure (as the schematic shown in Fig. 1(c)) with CdSe/ZnS QD and WSe₂ nanoflakes. PL enhancement with the EF value of 3.5 has been determined with the thickness of WSe₂ at about 40 nm and that of QD at about 100 nm, as shown in Figs. 5(c) and 5(d), respectively (see Fig. S18 for details). For the heterostructures, such as h-BN/WSe₂ heterostructures shown in Fig. 5(e), PL modulation can also be determined. Again, PL intensity in the QD-h-BN hybrid structures generally presents weak enhancement (see Fig. S19), consistent with the predictions shown in Fig. 1(j). The comprehensive experiments definitely reveal the success of our multiple reflections model and the presence of optical interference effect in the different hybrid QD-2D structures. By optimizing the relevant parameters, the PL intensity of the QD films can be well enhanced, and thus the optical performance of QD-based devices can be further improved by designing the applicable hybrid structures.

 

Fig. 5. PL modulation of the QD films under different hybrid structures. (a) PL image of the hybrid A/(CdSeTe/ZnS)/WSe₂/SiO₂/Si structure with the QD films capping on the WSe₂ nanoflakes. The inset presents the optical image of the WSe₂ nanoflakes. The thicknesses of WSe₂ marked by the solid line are 15.3 nm and 29.1 nm, respectively (see Fig. S17 for details). (b) PL intensity of the selected line. (c) PL image of the hybrid G/(CdSe/ZnS)/WSe₂/A structure, the inset presents the corresponding optical image. The thickness of the WSe₂ is 41.8 nm (see Fig. S18 for details). (d) PL intensity of the selected line. (e) PL image of the QD films capping the h-BN/WSe₂ heterostructure. The inset shows the optical image of the heterostructure, with the h-BN covering the WSe₂. The thicknesses of the h-BN and WSe₂ are 51.5 nm and 19.1 nm, respectively (see Fig. S19 for details). (f) PL intensity of the se lected line. The averaged intensity with the standard derivations of each level has been depicted by the color shadows. Scale bars: 40 µm.

Download Full Size | PDF

3. Conclusions

Here we theoretically and experimentally study the interactions between the QD films and 2D materials, by measuring the PL intensity and spectra of the QD films. For thin QD films and/or few-layer 2D materials, energy transfer and electron transfer from the QD to 2D materials will dramatically quench the PL of QD. With the increase of the QD films and 2D materials, the optical interference effect originating from the multiple reflections in the interface between the QD films and 2D materials will enhance the PL intensity of the QD films. The effect of PL enhancement can be periodically modulated by changing the thicknesses and the kinds of both QD and 2D materials, as well as the wavelength of the excitation and emission lights. We have proved the effectiveness of our physical model by performing PL measurements of two kinds of QD on different hybrid QD-2D structures. In the hybrid A/(CdSe/ZnS)/WSe₂/SiO₂/Si structure, PL intensity can be altered from almost disappearing (EF∼0.08) to strong enhancement (EF∼6), and the experimental EF are in good agreement with the theoretical calculations. PL modulations have also been determined with the other 2D materials (such as MoS₂ and h-BN) and QD (CdSeTe/ZnS). The optical interference effect uncovered in this work enables a powerful method to manipulate the PL property of the QD films in the different QD-2D hybrid structures. These results can boost the optical performance of the QD-based electronic and optoelectronic devices in the hybrid QD-2D structures.

4. Methods

4.1 Sample preparation and characterization

Water-soluble CdSe/ZnS with an emission peak at 625 nm was purchased from Mesolight Inc (Suzhou Co. Ltd). The near-infrared CdSeTe/ZnS core-shell QD with the emission peak at 800 nm (Qdot 800ITKTM Organic Quantum Dots) was ordered from Thermo Fisher Scientific. The PL QY of both QD is larger than 80%. 50 µL original QD with a concentration of about 10 nMol/mL was further diluted by deionized water (1 mL) and water-soluble PVA solution (1 mL with the concentration of 40 mg/mL). Typically, the mixture was spin-coated on the substrates with a rotational speed of 600 rmp for 1 minute followed by 2000 rmp for 2 minutes.

TMDC materials and h-BN nanoflakes with different thicknesses were prepared on the SiO₂/Si substrates by mechanical exfoliation from the bulk crystals, purchased from SixCarbon Technology (Shenzhen, China). For the A/QD/2D/SiO₂/Si hybrid structures, the final samples were prepared by spin-coating QD mixture on the SiO₂/Si substrates with these 2D nanoflakes. The thickness of the QD films was controlled by optimizing the spin-coating parameters. For the G/QD/2D/A hybrid structures, the QD films were firstly prepared on the clean glass coverslips by spin-coating the mixtures, and then transferred the fabricated nanoflakes on the QD films by the dry transfer method. We have performed PL measurements with and without the tapes (Gelfilm from Gelpak), their influence on the multi-layer interference effect for the hybrid G/QD/2D/A structures can be ignored (see Fig. S12). The thickness of the prepared samples, including 2D nanoflakes and the QD films, was defined by AFM (FlexAFM C3000, Nanosurf Inc.). Raman spectra were performed by a LabRAM HR Raman microscope (Horiba Inc.). The morphology of the final samples was also characterized by an optical microscope (BX53M, Olympus) coupled with a CCD (SC50, Olympus).

4.2 Optical setup for PL measurements

The PL measurements were performed using a home-built scanning confocal system based on an inverted microscope (TE2000-U, Nikon). The experimental setup is sketched in Fig. S20 and the detailed descriptions can be found in our previous works [41,42]. Particularly, a continuous-wave laser with a wavelength of 532 nm was used to excite both QD and TMDC nanoflakes. The hybrid structures were placed on an XY piezo scanner (PXY 102 SG, PiezoSystem Jena Inc.). PL imaging was achieved by moving the samples with respect to the laser spot in a controlled way. The laser beam was focused by a 100× objective with a numerical aperture (NA, Nikon) of 0.9. PL from QD and/or TMDC nanoflakes were collected with the same objective. After passing through a dichroic mirror (Di02-R532-25×36, Semrock Inc.) and a notch filter, PL was split into two beams by a beam splitter with a ratio of 70:30. The stronger beam was collected by a monochromator equipped with a cooled CCD (PIXIS, Princeton Instrument Inc.) to perform PL spectra. The weaker beam was further filtered spatially by a 100 µm pinhole and split into two equivalent beams by another beam splitter with a ratio of 50:50. PL intensity of QD was filtered by a narrow emission filter (FF01-640/40-25, Semrock Inc.) and detected by a single-photon counting modulator (SPCM-AQR-15, PerkinElmer Inc.). On the other hand, PL of TMDC was filtered by another narrow band filter.