Highly Efficient Upconversion Emission Platform Based on the MDM Cavity Effect in Aluminum Nanopillar Metasurface

Abstract

Rare earth-doped upconversion nanoparticles (UCNPs) can convert low-energy photons (NIRs) into high-energy photons (visible light), offering advantages such as low background signal, good stability, and excellent biocompatibility. However, exploring a strategy to combine the advantages of high efficiency, low cost, and easy fabrication of a plasmonics–UCNPs system is still a challenge. Here, we reported a metal–dielectric–metal (MDM)-type plasmonic platform based on the aluminum metasurface, which can efficiently enhance the luminescence intensity of magnetic and non-magnetic rare earth-doped UCNPs. Attributed to the strong local field effect of the nanocavities formed by the aluminum anti-transmission layer at the bottom, the fluorescence of the two types of UCNPs in such a platform can be enhanced by over 1000 folds compared with that in the conventional substrate. It is found that the deposited UCNPs amount and the aluminum pillar size can both impact the enhancement. We confirmed that the constructed MDM nanocavities could enhance and regulate the local field strength, and the optimum enhancement can be achieved by choosing proper parameters. All these findings provide an efficient way of exploring the plasmon-enhanced UCNPs luminescence system with low cost, high efficiency, and easy fabrication and can be promising in the fields of biosensing and photovoltaic devices.

1. Introduction

In recent years, lanthanide-doped upconversion nanoparticles (UCNPs) have attracted considerable attention due to their ability to convert low-energy near-infrared light into high-energy visible or ultraviolet emission [1,2,3,4]. They can emit high-energy short-wavelength light under low-energy excitation photons. This characteristic makes UCNPs promising in biomedical imaging [5,6], photocatalysis [7], information storage, and encryption technology [8]. However, their low luminescence quantum yield [9,10,11,12] hindered their real-world applications.

To overcome this challenge, a series of strategies have been applied to enhance the emission efficiency of UCNPs [13,14,15]. The idea of combining UCNPs with a variety of plasmonic nanostructures was proved to be efficient [16,17,18,19]. For example, Wu et al. designed a plasmonic nanocavity structure composed of silver nanocubes and a gold mirror [18], which couples the radiative upconversion emission into nanophotonic hotspots with strongly enhanced local electromagnetic fields, resulting in a remarkable enhancement of nonlinear upconversion luminescence by a factor of 1 × 10⁴ under low-power excitation. Despite these advancements, most plasmonic structures still suffered the problem of scattering, reflection, or transmission loss of the excitation power due to the weak absorbance. Compared with the common plasmonic structures such as grating, hole, or pillar arrays, the sandwiched metal–dielectric–metal (MDM) structures can result in nearly perfect absorption of the input light with certain wavelength, which can promote the UCNPs emission [20,21,22,23,24]. Sun et al. used optical lithography to fabricate pyramid-shaped grating plasmonic substrates [22], successfully increasing the resonant energy transfer rate from Yb³⁺ to Er³⁺ 5.1 times. Additionally, Das et al. constructed a gold–UCNPs–gold sandwich structure [24] using lithography, achieving over a thousand-fold enhancement in photoluminescence intensity and realizing a resonant absorption of 980 nm of light through geometric optimization.

However, traditional MDM plasmonic arrays require the nanolithography technique, such as photolithography or electron beam lithography. Moreover, common MDM structures usually applied noble metal such as gold and silver as the plasmonic platform. The high cost in fabrication process and materials limited the large-scale and economical production of efficient upconversion platforms [25,26]. Therefore, finding alternative materials has become a research priority. In recent years, the aluminum (Al)-based nanostructure has gained significant attention as a low-cost material with plasmonic effects [27,28,29,30,31]. Liu et al. embedded UCNPs into the gap plasmonic resonator structure of an aluminum nanodisk array [27], utilizing the gap plasmon mode in the MIM nanostructure to enhance the luminescence intensity of UCNPs by two orders of magnitude. As the UCNPs-based applications extended from the biosystems to energy area, UCNPs with different excitation wavelengths and various properties should be considered. To achieve this, the plasmonic structure should be flexible in tuning its surface plasmon resonance (SPR) property for Nd³⁺- and Yb³⁺-sensitized UCNPs, or the magnetic and non-magnetic UCNPs. Therefore, developing a compatible plasmonic upconversion platform that is simple to fabricate, tunable in performance, low cost, and widely applicable is of great importance.

To address this, we propose an economic strategy using the nanoporous anodic aluminum oxide [32,33] (AAO) template to fabricate aluminum nanoplasmonic structures. This approach enhances upconversion luminescence while simplifying the fabrication process and reducing production costs. We systematically investigated the effects of structural parameters (e.g., pillar radius and height) on plasmonic absorption peaks across different UCNPs material systems, aiming to achieve significant improvements in luminescence efficiency. The bottom reflector can reduce the original transmission loss and partly reuse the losing light, which eventually enhances the upconversion emission. The results demonstrated that this strategy is not only economic and efficient but also compatible for the magnetic (Gd³⁺ doping) and non-magnetic UCNPs. The low cost in fabrication process and materials will promote the large-scale and economical production of efficient upconversion platforms, which will boom its applications in bioimaging, optoelectronics, and solar cells.

2. Structure and Model

Here, we firstly demonstrate the commonly used plasmonic structure without the bottom metal reflector, where the upconversion nanoparticles (UCNPs) layer is directly deposited on the aluminum array, as shown in Figure 1a. The related absorption spectra and electric field distribution at 815 nm are shown in Figure 1c. Due to significant transmission and reflection losses, this design exhibits limited absorption efficiency and local electric field enhancement. To address this, a modified structure with a bottom aluminum layer coated under the silicon dioxide layer is proposed, shown in Figure 1d. This structure consists of a 100 nm thick aluminum reflective layer, a silica (SiO₂) substrate with a depth of d, a UCNPs layer with a thickness of H, and a top aluminum array. The detailed geometric parameters of the aluminum array are shown in Figure 1e, where the height and radius of the cylindrical units are h and r, respectively, and the period P = 300 nm. This optimized design effectively enhances the surface plasmon resonance (SPR) effect [34,35], with the underlying mechanism illustrated in Figure 1b. Due to the reflection of the bottom Al layer, the originally transmitted light was reflected and re-localized by the upper plasmonic pattern. This will eventually boom the emission intensity of the UCNPs. It can be found that after the bottom layer was deposited, the absorption efficiency around 808 nm was more than doubled.

Such a plasmonic structure can be fabricated by the nanoporous anodic aluminum oxide (AAO) template approach, as shown in Figure 2a. First, a flat and clean silicon dioxide (SiO₂) layer is deposited on the aluminum layer. Then, an AAO template is covered on top, and aluminum is deposited into the pores of the AAO template to form the nanopillar structure [36]. Next, the AAO template is removed using a phosphoric acid solution, leaving behind an ordered aluminum nanopillar array [35]. Finally, an upconversion nanoparticle (UCNPs) layer is deposited onto the structure to complete the plasmon-enhanced configuration. The method offers advantages such as low cost, simple operation, and good structural uniformity. Therefore, from the perspective of experimental implementation, it provides a feasible and effective approach for the practical fabrication of such complex plasmonic structures. To investigate the optical properties of this complex structure, we employed the three-dimensional finite-difference time-domain [37] (FDTD) method for numerical simulations. In the simulation setup, periodic boundary conditions were applied in the x and y directions, while perfectly matched layer (PML) boundary conditions were used in the z direction to ensure the accuracy of the simulation results. The excitation light was injected along the z direction. Notably, the aluminum nanoplasmonic array structure exhibits a significant absorption peak at 808 nm, as shown in Figure 1c, which matches the excitation band of the sensitizer ion Nd³⁺ in the upconversion nanoparticles (UCNPs). This spectral overlap [38] indicates the presence of SPR at this wavelength, satisfying a key condition for plasmon-enhanced upconversion. The plasmon-enhanced upconversion mechanism is illustrated in Figure 2b: the plasmon-enhanced absorption increases the electron population in the excited state of Nd³⁺ ions and, via energy transfer, also enhances the excited-state population of neighboring Yb³⁺ ions. This subsequently promotes the population of the excited states in activator ions such as Tm³⁺, ultimately leading to enhanced upconversion emission.

3. Results and Discussion

For numerical simulations, NaYF₄ was selected as the host material for upconversion nanoparticles, and the refractive index (n) of NaYF₄ can be approximately set as 1.42. The band overlap between plasmonic structures and UCNPs is critical for amplifying upconverted emission, prompting a systematic study of the plasmonic properties [39,40,41]. Key structural parameters were varied to explore the enhancement mechanism and optimize the design.

Firstly, the UCNPs layer thickness (H) was varied from 280 nm to 400 nm with a step size of 20 nm, while keeping the SiO₂ layer thickness (d) of 100 nm, the pillar radius (r) of 100 nm, and the pillar height (h) of 300 nm constant. As shown in Figure 3a, the absorption peak redshifts as H increases, reaching a local absorption maximum at H = 320 nm. Notably, when H increases to 360 nm, the absorption spectrum extends to 808 nm, which matched well with the excitation band of the Nd³⁺-sensitized UCNPs. Figure 3b shows the electric field intensity profile (|E/E₀|²) at 808 nm for different H values (x–z plane). This thickness-sensitive local field intensity can also affect the light–matter interaction in turn. It can be found that the UCNPs with thickness H = 360 nm can be efficiently excited and enhanced by the local field at wavelength 808 nm. This indicated that properly loaded UCNPs can maximize the enhancement factor.

The Al pillar height (h) can also be effective in tuning the absorption spectra. Here, we changed the pillar height from 300 to 400 nm with a step of 20 nm while keeping the other parameters constant. As the pillar height (h) increased, the absorption peak redshifted and the efficiency weakened, as shown in Figure 3c. Notably, when h reached 320 nm, the absorption peak was precisely located at 808 nm. This can be explained by the SPR mode in the Al pillar. As the pillar height (h) increases, the aspect ratio (i.e., the ratio of height to diameter) of the pillar also increases [42,43,44]. This lengthens the oscillation path of electrons in the height direction, thereby reducing the resonance frequency of the longitudinal SPR. To visually demonstrate this effect, Figure 3d showed the field distribution in the x–z plane at the 808 nm wavelength for different pillar heights (h = 300 nm, 320 nm, 340 nm, and 360 nm). Notably, when h = 320 nm, the local electric field at this wavelength reached a maximum value of |E/E₀|ₘₐₓ = 8, indicating a significant local field enhancement.

Besides the deposited UCNPs thickness and the height of the Al pillar, the absorption was also sensitive to the pillar radius r. The relevant absorption spectra with radius r from 60 nm to 130 nm are shown in Figure 4a. A clear absorption enhancement was found as r increased. This can be explained by the improved field localization capability when the pillar got thicker. As a result, the absorption efficiency was enhanced as r increased. In addition, the normalized field strength ∣E/E₀∣ at 808 nm was also gradually enhanced (Figure 4b). However, when r was enlarged up to 130 nm, both the absorption peak intensity and normalized field strength at 808 nm sharply dropped.

Finally, we focused on the SiO₂ layer depth d, which played a critical role in enhancing the absorption and reduced the transmitted loss. In this structure, the SiO₂ layer is sandwiched between the bottom metallic reflector and the top aluminum pillar array, forming a Fabry–Pérot-like plasmonic resonant cavity [45]. As a dielectric spacer between two metallic components, the SiO₂ layer meets the resonance conditions necessary for localized electromagnetic field enhancement, thereby exhibiting the typical characteristics of a plasmonic cavity. By increasing the cavity depth from 40 nm to 260 nm while keeping the UCNPs layer thickness (H) of 400 nm, the pillar height (h) of 300 nm, and the pillar radius (r) of 100 nm constant, the obtained absorption spectra are shown in Figure 4c. The plasmonic cavity size enlarged as d increased; thus, the resonant wavelength showed a gradual redshift. The slightly weakened absorption efficiency can be attributed to the mode transformation from cavity mode (d = 60 nm) to dipole radiation mode (d = 260 nm). The evidence was that most of the input power was localized at the SiO₂ layer for d = 60 nm while being distributed on the upper surface of the Al pillar for d = 260 nm (Figure 4d). The local electric field intensity ∣E/E₀∣ decreased from 7.92 to 6.88 as the cavity size increased. Therefore, choosing proper SiO₂ layer thickness can create a balance between the resonant wavelength and the local field intensity.

This plasmonic platform is not only suitable for the above non-magnetic UCNPs (NaYF₄ host) but also extended to application with magnetic matrix NaGdF₄-based UCNPs. Here, the refractive index of typical magnetic UCNPs with a NaGdF₄ host was chosen to be n = 1.89 [46]. When the UCNPs matrix material was replaced with magnetic NaGdF₄, under fixed parameters (SiO₂ layer depth d = 100 nm, Al pillar radius r = 100 nm, height h = 300 nm), while the UCNPs layer height H increased from 240 nm to 360 nm, a similar spectra evolution can be observed (Figure 5a). And the absorption peak continuously redshifts as H increases, while the peak intensity shows a non-monotonic change, first increasing and then decreasing. Notably, at H = 263 nm, the absorption peak is located at the 808 nm wavelength, with an absorption peak of 95%. This indicates that the replacement of the matrix material not only affects the absorption peak but also enables the tuning of the SPR wavelength.

For the plasmonic platform with a magnetic NaGdF₄ host matrix, Figure 5b,c show the influence of nanopillar radius r and height h on the absorption spectra. It can be found that enlarging the pillar radius can effectively improve light harvesting capabilities and strengthen light–matter interactions (as shown in Figure 5b). In contrast, the absorption peak redshifts with increasing h, while the peak intensity gradually weakens, as shown in Figure 5c. The SiO₂ thickness-dependent absorption is analyzed and shown in Figure 5d. The plasmonic cavity size enlarged as d increased; thus, the resonant wavelength showed a gradual redshift. The slight decrease in absorption efficiency can be attributed to the transition from a strongly coupled cavity mode at a smaller gap (d = 40 nm) to a dipole radiation-dominated mode at a larger gap (d = 260 nm). At smaller gaps, the localized surface plasmon resonance is confined within the nanogap, whereas at larger gaps, interparticle coupling weakens and radiative losses increase [47]. Based on the results above, such a plasmonic platform shows potential applicability to both magnetic and non-magnetic UCNPs.

In the plasmon-enhanced upconversion process, the excitation field and the decay rate of the intermediate state are the main determining factors. Therefore, the total upconversion enhancement factor (EF_upc) is defined as the product of the excitation enhancement factor (EF_exc) and the QY enhancement factor (EF_QY) as EF_upc = EF_exc × EF_QY. Under conditions where the SPR wavelength is matched with the excitation band of the UCNPs, EF_exc is typically dominates over EF_QY [48]; thus, EF_upc ≈ EF_exc. Recent studies [49,50] have shown that EF_exc is approximately proportional to the fourth power of the local electric field enhancement (|E/E₀|), leading to the approximation: EF_upc ≈ |E/E₀|⁴, where |E/E₀| represents the volume-averaged local electric field enhancement within the UCNPs layer. The best performance comparison of the plasmonic structure with magnetic (NaGdF₄) and non-magnetic (NaYF₄) UCNPs are shown in Figure 5e. For UCNPs with the NaYF₄ matrix, the combination of H = 310 nm, r = 110 nm, h = 320 nm, and d = 100 nm resulted in an absorption peak of 96.12%. For UCNPs with the NaGdF₄ matrix, the combination of H = 268 nm, r = 110 nm, h = 260 nm, and d = 100 nm achieved an absorption peak of 99.94%. Although the NaGdF₄ matrix exhibits higher absorption efficiency, the NaYF₄ matrix demonstrates stronger local field intensity, with a total upconversion enhancement factor of 2.89×10³, which is 50% higher than that of the NaGdF₄ system. This contradiction arises from differences in material properties: the lower refractive index of NaYF₄ promotes the strong localization of SPR, while the higher refractive index of NaGdF₄ enhances light trapping efficiency. Thus, these AAO template-assisted plasmonic platforms can be effective in enhancing the upconversion efficiency of both magnetic (NaGdF₄) and non-magnetic (NaYF₄) UCNPs as the enhancement factor for these two structures was over 1.8 × 10³.

Finally, the local electric field distributions at 808 nm for these two structures were discussed to evaluate the photon upconversion efficiency enhancement mechanism, as shown in Figure 5f. The local electric field intensity in the x–y plane for the NaYF₄ matrix structure reached 8.76, which is higher than the value of 7.1 obtained for the NaGdF₄ matrix structure. This strong local field enhancement not only improves the interaction between photons and matter but also further optimizes the upconversion luminescence efficiency. In summary, the choice of matrix material plays a critical role in optimizing the performance of UCNPs structures. The lower refractive index of NaYF₄ resulted in a stronger local field, while the higher refractive index of NaGdF₄ enhanced light absorption efficiency.

4. Conclusions

In conclusion, this study proposed a highly efficient upconversion emission platform based on the MDM cavity effect in an aluminum nanopillar metasurface. The upconversion enhancement factor was found to be sensitive to the deposited UCNPs amount and the aluminum pillar size. And the constructed MDM nanocavities were proved to be efficient and dominant in regulating the local field strength and distribution. The results demonstrate that NaYF₄-based UCNPs exhibit optimal performance under 808 nm excitation, achieving an upconversion luminescence enhancement factor of 2.89 × 10³. And such a plasmonic platform was useful to both the magnetic (NaGdF₄-hosted) and the non-magnetic (NaYF₄-hosted) UCNPs. All these findings provide an efficient way of exploring the plasmon-enhanced UCNPs luminescence system with low cost, high efficiency, and easy fabrication and can be promising in the fields of bioimaging, optoelectronics, and solar cells.