Near‐Infrared‐Plasmonic Energy Upconversion in a Nonmetallic Heterostructure for Efficient H2 Evolution from Ammonia Borane

Abstract Plasmonic metal nanostructures have been widely used to enhance the upconversion efficiency of the near‐infrared (NIR) photons into the visible region via the localized surface plasmon resonance (LSPR) effect. However, the direct utilization of low‐cost nonmetallic semiconductors to both concentrate and transfer the NIR‐plasmonic energy in the upconversion system remains a significant challenge. Here, a fascinating process of NIR‐plasmonic energy upconversion in Yb3+/Er3+‐doped NaYF4 nanoparticles (NaYF4:Yb‐Er NPs)/W18O49 nanowires (NWs) heterostructures, which can selectively enhance the upconversion luminescence by two orders of magnitude, is demonstrated. Combined with theoretical calculations, it is proposed that the NIR‐excited LSPR of W18O49 NWs is the primary reason for the enhanced upconversion luminescence of NaYF4:Yb‐Er NPs. Meanwhile, this plasmon‐enhanced upconversion luminescence can be partly absorbed by the W18O49 NWs to re‐excite its higher energy LSPR, thus leading to the selective enhancement of upconversion luminescence for the NaYF4:Yb‐Er/W18O49 heterostructures. More importantly, based on this process of plasmonic energy transfer, an NIR‐driven catalyst of NaYF4:Yb‐Er NPs@W18O49 NWs quasi‐core/shell heterostructure, which exhibits a ≈35‐fold increase in the catalytic H2 evolution from ammonia borane (BH3NH3) is designed and synthesized. This work provides insight on the development of nonmetallic plasmon‐sensitized optical materials that can potentially be applied in photocatalysis, optoelectronic, and photovoltaic devices.


Experimental Section
(1) Synthesis of Upconversion Luminescence Nanoparticles (NPs): 0.78 mmol of YCl 3 ⋅6H 2 O, 0.20 mmol of YbCl 3 ⋅6H 2 O, and 0.02 mmol of ErCl 3 ⋅6H 2 O were mixed with a solution consisting of 6 mL oleic acid and 15 mL octadecene in a 100 mL flask, which was then heated to 160°C for 30 min under N 2 gas protection to form a homogeneous solution. Subsequently, this solution was cooled to room temperature at a flow of N 2 gas though the reaction flask. During this process, 10 mL of methanol solution containing NaOH (2.5 mmol) and NH 4 F (4 mmol) was slowly dropped into the reaction flask. The obtained solution was stirred for 30 min. Subsequently, the reaction temperature was increased to 80°C to evaporate methanol from the reaction solution; in succession, the solution was heated to 300°C where it was maintained for 1 h under N 2 atmosphere. When the solution had naturally cooled down, the products ofNaYF 4 :Yb-Er NPs with mean sizes of 40 nm were collected via centrifugation, and washed with ethanol/cyclohexane (3:1 v/v) thrice and finally dispersed in cyclohexane solvent with a concentration of ~0.1 M. TheNaYF 4 :Yb-Er NPs with mean sizes of 10 nm and 20 nm were obtained through controlling the reaction temperature at 280 and 290°C, respectively. Please note that when the reaction temperature was reduced to 280°C, the 10-nm NaYF 4 :Yb-Er NPs could be obtained while their phase structures changed from  to .
(2) Fabrication of films of plasmonic W 18 O 49 Nanowires (NWs) on FTO glass: In a typical process, 25 mg of W(CO) 6 was dissolved into 20 mL of absolute ethanol under constant stirring to form a yellow transparent solution. Then, a cleansed FTO glass at a size of 2 cm 3 cm was placed into a Teflon-lined autoclave, loaded with the above solution. The autoclave was sealed and then maintained at 180°C for 12 h. The obtained sample on the FTO glass with blue color was removed from the reaction solution, washed with ethanol, and finally dried in nitrogen.
(3) Fabrication of NaYF 4 /W 18 O 49 films: In a typical process, 0.1 mL of NaYF 4 :Yb-Er NPs-suspended cyclohexane solution (0.1 M) was dropped into 10 mL of cyclohexane solution under ultrasonic treatment for 20 min. After that, the FTO glass loaded with plasmonic W 18 O 49 NWs film was vertically immersed into the above suspended solution and then put in an oven at 35°C for 2 h. With the slow evaporation of ethanol solvent, a thin film consisting of NaYF 4 :Yb-Er NPs laxly self-assembled onto the surface of W 18 O 49 NWs grown on the FTO glass. Based on a similar fabrication method, two control samples were also constructed on the FTO glass: one was the individual NaYF 4 :Yb-Er film that was obtained via direct self-assembly of the  Figure S1).
(4) Synthesis of NaYF 4 :Yb-Er@W 18 O 49 quasi-core/shell heterostructures: For the first step, 1.2 mmol of NaCl, 0.48 mmol of YCl 3 , 0.108 mmol of YbCl 3 , and 0.012 mmol of ErCl 3 were mixed in 9 mL of ethylene glycol (EG) solvent to form a transparent solution that was labeled as solution A. Concurrently, 3.0 mmol of NH 4 F and 0.006 mmol of polyethyleneimine were dissolved into 6 mL of EG solvent to obtain the other transparent solution that was labeled solution B. The resulting solutions A and B were then mixed and agitated for 10 min, transferred into a 25 mL of Teflon-lined autoclave, and kept at 200°C for 2 h. The products of hydrophilic -NaYF 4 :Yb-Er NPs with a mean size of 40 nm were collected via centrifugation, washed with ethanol four times and finally dispersed in an ethanol solvent at a concentration of ~0.1 M; For the second step, 25 mg of W(CO) 6 was dissolved into 20 mL of absolute ethanol under constant stirring. Then, 50 L of the -NaYF 4 :Yb-Er NPs(40 nm)-suspended ethanol solution (0.1 M) was gradually dropped in the above solution and sealed in a Teflon-lined autoclave. The autoclave was maintained at 180°C for 12 h. The precipitates with blue color were collected via centrifugation, washed with ethanol four times, and finally dried in vacuum.

Catalytic H 2 evolution:
2 mg of the as-synthesized catalysts were dispersed into 4 mL of deionized water (Scheme S1). The obtained suspension solution was placed in a photoreactor with a volume of 35 mL under constant stirring. The reactor was then sealed and degassed with argon for 10 min. Subsequently, 2 mL of deionized water containing 2 mg of ammoniaborane was injected into the photoreactor, which was exposed to 980-nm laser diode or a 300 W Xe lamp (PLS-SXE300UV) coupled with a monochromator (The intensity of monochromatic light at ~8 mw/cm 2 ). The surrounding temperature was fixed at 28℃ by using a connected reflux water condenser. The generated H 2 was periodically analyzed via gas chromatograph, equipped with a thermal conductivity detector (Beifen-Ruili Analytical Instrument, SP-3420A).
Scheme S1 Schematic illustration of the photocatalytic experimental setup.

Characterization
X-ray diffraction (XRD) patterns of the as-synthesized samples were recorded via Shimadzu XRD-6000 X-ray diffractometer with a Cu Kα line of 0.1541 nm. Scanning electron microscopy (SEM; XL-30 ESEM FEG, Micro FEI Philips) and transmission electron microscopy (TEM; JEOL JEM-2100) were used to investigate the morphologies and structures of samples. The UV-vis absorption spectra of the samples were recorded on a Lambda 750 UV-Vis-NIR spectrophotometer (Perkin-Elmer, USA). The upconversion emission properties of products were measured with an inverted microscope (Olympus IX71) combined with a spectrometer (PI Instrument). Excitation with a 980-nm laser passing a laser clean-up was reflected into the objective (50X Olympus) via a dichroic short pass filter. The upconversion emission was collected with the same objective and conducted into the spectrometer. The laser clean-up and dichroic filter was used to purify the excitation light and to eliminate the laser line before the emission was detected (Scheme 1).
Scheme S2 Schematic illustration of the set-up used to measure the upconversion emissions of as-fabricated products.

Finite element method (FEM) simulation
All full wave numerical simulations were performed with the finite element method (FEM, commercial software package, Comsol Multiphysics 5.0). The W 18 O 49 (permittivity was obtained with the following simulation. See scheme S2) nanowire (diameter D 1 = 10 nm, length L = 800 nm) bundle (three wires touching each other) was placed in a homogeneous surrounding medium with an effective refractive index of 1.0. NaYF 4 sphere (permittivity ε = 2.477 at 980 nm, diameter D 2 = 40 nm) was put in direct contact with the W 18 O 49 nanowire at different positions. Non-uniform meshes were used to format the object. The largest mesh was set to below /6  . A perfect matched layer (PML) was used to minimize scattering from the outer boundary. The structure was placed in the x-y plane. The incident light was set to 1 V/m polarized in y-direction and propagated in z-direction.

Optical property simulation for W 18 O 49
The Cambridge Serial Total Energy Package (CASTEP) has been used for optical property calculations, which is based on the density functional theory (DFT) and utilizes a plane-wave pseudopotential method. [1] We used the generalized gradient approximation (GGA) in the scheme of Perdew-Burke-Ernzerhof (PBE) to describe the exchange-correlation functional. [2] The interaction between valence electrons and the ionic core was described via ultrasoft pseudopotential. [3] An energy cutoff of 300 eV was chosen for the W 18 O 49 crystal. The Brillouin-zone sampling mesh parameters for the k-point set were 2×2×2 for 67 atoms systems. In the optimization process, the energy change, maximum force, maximum stress, and maximum displacement tolerances were set as 2×10 −6 eV/atom, 0.05 eV/Å, 0.1 Ga, and 0.002 Å, respectively.        We carried out the time-resolved luminescence spectroscopy measurements, which is particularly useful for probing the actual influence of emission-matched plasmonic nanostructures on upconversion NPs. In our case, the lifetimes of 2 I 11/2  4 I 15/2 (521 nm), 4 S 3/2  4 I 15/2 (545 nm), and 4 F 9/2 -4 I 15/2 (660 nm) decays for the NaYF 4 :Yb-Er/W 18 O 49 film were shorter than the corresponding lifetimes obtained from the individual NaYF 4 :Yb-Er film ( Figure S6 a-f). However, the comparison of upconversion luminescence spectra between the NaYF 4 :Yb-Er and NaYF 4 :Yb-Er/W 18 O 49 films indicated that the introduction of W 18 O 49 NWs into the NaYF 4 :Yb-Er film could lead to a remarkable enhancement of green emission from the 2 I 11/2  4 I 15/2 (521 nm) transition, but the quenched emissions from both 4 S 3/2  4 I 15/2 (545 nm) and 4 F 9/2 -4 I 15/2 (660 nm) transitions. A decrease in decay lifetime coincident with an increase in emission suggests the enhancement of radiative rate of the 2 I 11/2  4 I 15/2 (521 nm) transition. On the other hand, the decreased decay lifetime and emission quenching indicate the enhanced non-radiative decay process for the 4 S 3/2  4 I 15/2 (545 nm) and 4 F 9/2 -4 I 15/2 (660 nm) transitions. [4] These results reveal the existence of plasmon-mediated competition between the radiative and non-radiative processes in the NaYF 4 :Yb-Er/W 18 O 49 film after 980-nm excitation. Figure S6g shows the pump power dependence of red and green upconversion emission intensity of NaYF 4 :Yb-Er NPs. After a linear fitting analysis, we found that the slopes are 1.95 and 1.59 for red and green emissions. The values are ~2, indicating the two-photon required for the upconversion process.  To understand qualitatively the photophysical mechanism in this plasmon-enhanced upconversion luminescence process, a set of rate equations were established based on the well -known upconversion process:    N 5 , and N 6 denote the population densities of 4 I 15/2 , 4 I 13/2 , 4 I 11/2 , 4 I 9/2 , 4 F 9/2 , ( 2 H 11/2 + 4 S 3/2 ), and 4 G 11/2 levels of Er 3+ , respectively. N Yb0 , and N Yb1 are the population densities of 4 F 7/2 and 4 F 5/2 levels of Yb 3+ , respectively. N Er and N Yb are the nominal ions densities corresponding to Er and Yb codoping concentrations, respectively. R 1 , 4 , R 5 , and R 6 are the radiative rates of 4 I 13/2 , 4 I 11/2 , 4 I 9/2 , 4 F 9/2 , ( 2 H 11/2 + 4 S 3/2 ) and 4 G 11/2 levels of Er 3+ , respectively. R ij ' is non-radiative rate from level i to level j. R G '' and R R " are the non-radiative energy transfer rates from the ( 2 H 11/2 + 4 S 3/2 ) and 4 /dt=0), the rate equations can be simplified as:

dN R N C N N R N R N dt
The intensities of red (I Red ) and green (I Green ) emissions can be given according to the population densities of N 4 From Equation (18) and (19), we can see that both the I Red and I Green values increase with the LSPR-enhanced excitation (f ex ) and emission (f 4 em and f 5 em ) fields. However, the transfer process of non-radiative energy from the excited state (R G '' or R R ") of NaYF 4 :Yb-Er NP to the neighboring W 18 O 49 NW leads to the decrease of upconversion luminescence through the competition with the LSPR-enhanced emission field . Please note that this non-radiative energy transfer is very sensitive to the interaction distance between the luminescent center of NaYF 4 :Yb-Er (Er 3+ ion) and the plasmonic W 18 O 49 . Meanwhile, the energy transfer rate is related to the absorption intensity of W 18 O 49 acceptor. Because the absorption intensity of W 18 O 49 NWs in the red light region is much stronger than that in the green light region, the value of R R " should be much larger than that of R G ''. The finite element method simulations indicated the similar enhancement factors on the LSPR-enhanced electric fields at green and red emissions. Thus, the I Red should be much smaller than the I Green for the NaYF 4 :Yb-Er/W 18 O 49 film due to the enhanced non-radiative process. Moreover, from Equation (18) According to the results of finite element method simulations, the enhancement factor at green emission field is similar to that at the red emission field (f 5  To confirm the hypothesis of NIR-plasmonic energy upconversion process, we synthesized another control sample labeled as the NaYF 4 :Yb-Er/W-W 18 O 49 film, in which the W-W 18 O 49 NWs only showed a weak LSPR band in NIR region. The W-W 18 O 49 NWs were obtained by reducing the solvothermal temperature to 160℃. [5] As shown in Figure S9 (Figure S9 d). The enhancement factors of upconversion luminescence for 4 I 11/2  4 F 7/2 (660 nm), 4 S 3/2  4 I 15/2 (545 nm), and 2 I 11/2  4 I 15/2 (521 nm) transitions are 1.8, 1.7, and 3.6, respectively (Figure S9 e). The enhancement factor for the 2 I 11/2  4 I 15/2 (3.6) transition is only twice higher than the factor for the 4 I 11/2  4 F 7/2 (1.8) transition. This ratio value is extremely lower than the corresponding value (90) obtained from the NaYF 4 :Yb-Er/W 18 O 49 film, indicating the lack of NIR-plasmonic energy upconversion process to selectively quench the upconversion luminescence of NaYF 4 :Yb-Er NPs. The low enhancement effect on the upconversion luminescence in this case is attributed to the weakened LSPR intensity of W-W 18 O 49 NWs.
In the case of NaYF 4 :Yb-Er, the energy separation of ~840 cm -1 can allow the 2 H 11/2 level to be populated from the 4 S 3/2 level by thermal excitation, and a quasi-thermal equilibrium forms between these two levels, resulting in the variation in the transitions of 2 I 11/2  4 I 15/2 (521 nm) and 4 S 3/2  4 I 15/2 (545 nm) at an increased temperature. The luminescence intensity ratio of the green upconversion emissions from the 2 I 11/2  4 I 15/2 (521 nm) and 4 S 3/2  4 I 15/2 (545 nm) transitions can be expressed as the following formula: [6] Where N, g, , and are the number of ions, the degeneracy, the angular frequency and the emission cross-section of luminescence transitions from the 2 I 11/2 and 4 S 3/2 levels to the 4 I 15/2 level, respectively. is the energy separation between 2 H 11/2 and 4 S 3/2 levels, k is the Boltzmann constant and T is the absolute temperature. and represent the integrated intensities of the transitions from the 2 I 11/2 (521 nm) and 4 S 3/2 (545 nm) levels to the 4 I 15/2 level, respectively. It can be seen that with the increase of the local temperature surrounding the NaYF 4 :Yb-Er NPs, the R value would increase considerably. In our case, the large ratio of the green upconversion emissions from the 2 I 11/2  4 I 15/2 (521 nm) and 4     The photocatalytic stability of NaYF 4 :Yb-Er@W 18 O 49 heterostructure was investigated under 980-nm irradiation. The result showed that after the first cycling use, the NaYF 4 :Yb-Er@W 18 O 49 heterostructures was almost inactive for the catalytic H 2 evolution from NH 3 BH 3 . The loss activity of NaYF 4 :Yb-Er@W 18 O 49 heterostructures can be ascribed to the quenched LSPR of W 18 O 49 due to the reduced electron density (or surface oxygen vacancy) during the catalytic H 2 evolution. However, it has been reported that the quenched LSPR of W 18 O 49 could be re-activated through an electrochemical treatment. [7] By using this method, we realized the reactivation of NaYF 4 :Yb-Er@W 18 O 49 heterostructures for H 2 evolution under 980-nm irradiation. These results also demonstrate that the LSPR-enhanced catalytic activity of NaYF 4 :Yb-Er@W 18 O 49 heterostructures for H 2 evolution arises mainly from the transfer of hot electron.