A New Up-conversion Material of Ho3+-Yb3+-Mg2+ Tri-doped TiO2 and Its Applications to Perovskite Solar Cells.

A new up-conversion nanomaterial of Ho3+-Yb3+-Mg2+ tri-doped TiO2 (UC-Mg-TiO2) was designed and synthesized with a sol-gel method. The UC-Mg-TiO2 presented enhanced up-conversion fluorescence by an addition of Mg2+. The UC-Mg-TiO2 was utilized to fabricate perovskite solar cells by forming a thin layer on the electron transfer layer. The results display that the power conversion efficiency of the solar cells based on the electron transfer layer with UC-Mg-TiO2 is improved to 16.3 from 15.2% for those without UC-Mg-TiO2. It is demonstrated that the synthesized UC-Mg-TiO2 can convert the near-infrared light to visible light that perovskite film can absorb to improve the power conversion efficiency of the devices.

Background

More attentions have been paid to the perovskite solar cells (PSCs) in the field of solar cells [1-5]. The power conversion efficiency (PCE) of the PSCs has been exceeding 22% within a few years [6]. However, the perovskite materials usually absorb the visible light whose wavelength is less than 800 nm, and more than half of the solar energy is not be utilized, especially in the region of near-infrared (NIR). To solve the issues, one of the effective methods is to apply the up-conversion nanomaterial to perovskite solar cells by converting the NIR light to visible light that the perovskite can utilize [7-9]. The beta-phase sodium yttrium fluoride (β-NaYF4) is commonly used as the host lattice for rare earth ions to prepare the up-conversion materials. While the β-NaYF4-based up-conversion materials are insulator, which is not beneficial for the electron transfer [ETL] [10].

Titanium dioxide (TiO2) nanocrystal with anatase phase is commonly used as the electron transfer material in the perovskite solar cells due to its suitable energy band structure, low cost, and long stability [11-13]. However, the energy band gap of TiO2 is large (3.2 eV), which hampers its applications. To improve the applications of TiO2 in visible light and near-infrared region, some methods were explored. One of the effective methods is doping TiO2 with metal or non-metal [14-16]. Yu et al. [17] demonstrated that Ho3+-Yb3+-F− doped TiO2 could convert NIR light to visible light that can be absorbed by the dye-sensitized solar cells (DSSCs). Zhang and co-authors [18] proved that Mg-doped TiO2 can change the Fermi energy level of TiO2 to enhance the performance of perovskite solar cells.

In this work, we are preferred to combine the rear earth ions (Ho3+ and Yb3+) and the metal ion (Mg2+) doped TiO2 together to synthesize a new material with enhanced up-conversion fluorescence. Our purpose is to explore how the addition of Mg2+ affect the up-conversion fluorescence of TiO2 and to apply the up-conversion nanomaterial of Ho3+-Yb3+-Mg2+ tri-doped TiO2 to perovskite solar cells. The results display that the addition of Mg2+ enhanced the up-conversion emission of TiO2, and the application of Ho3+-Yb3+-Mg2+ tri-doped TiO2 improved the PCE of PSCs to 16.3% from 15.2%.

Methods/Experimental

Materials

Formamidinium iodide (FAI), Methylamium bromide (MABr), Lead diiodide (PbI2), 2,2′,7,7′-Tetrakis-(N,N-di-p-methoxyphenylamine)-9,9′-spirobifluorene (Spiro-OMeTAD), and lead dibromide (PbBr2) were purchased from Xi’an Polymer Light Technology Corp. (China). The SnO2 colloid solution was purchased from Alfa Aesar (tin (IV) oxide). Dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF), 4-tert-butylpyridine (TBP), and lithium bis (trifluoromethanesulfonyl) imide (Li-TFSI) were purchased from Shanghai Aladdin Bio-Chem Technology Co., LTD (China).

Synthesis of Ho3+-Yb3+-Mg2+ Tri-doped TiO2

The up-conversion material of Ho3+-Yb3+-Mg2+ tri-doped TiO2 was synthesized with a reported method [19] with some modifications. Firstly, a Titanium tetrabutanolate was obtained by mixing acetylacetone (AcAc) and Titanium tetrabutanolate (Ti(OBu)4) for 1 h under stirring at 25 °C, and then the isopropyl alcohol (IPA) was added to prepare the (Ti(OBu)4) solution. A mixed solution of IPA, HNO3, and H2O was dropped into the solutions slowly. After stirring for 6 h, a TiO2 sol with a color of light yellow was obtained. In a typical synthesis, the molar ratio of AcAc, HNO3, and H2O to Ti(OBu)4 was 1:0.3:2:1. For the synthesis of Ho3+-Yb3+ co-doped TiO2, Ho(NO3)3·5H2O and Yb(NO3)3·5H2O were used as the elemental sources and added into the solution. Typically, the molar ratio of Ho3+:Yb3+:Ti = 1:x:100 (x = 2, 3, 4, 5). For the synthesis of Ho3+-Yb3+-Mg2+ tri-doped TiO2, Ho(NO3)3·5H2O, Yb(NO3)3·5H2O, and Mg(NO3)2 6H2O as the elemental sources were added into the solution, and the molar ratio of Ho3+:Yb3+:Mg2+:Ti = 1:4:x:100 (x = 0, 1, 1.5, 2, 2.5). The obtained solution was referred to as Ho3+-Yb3+-Mg2+ tri-doped TiO2 (UC-Mg-TiO2) sol. The solvent in the solution was removed by heating at 100 °C for 10 h. Then, the material powders were heated for 30 min at 500 °C.

Preparation of PSCs

The FTO was washed in detergent, acetone, and isopropanol, and then treated for 15 min with UV-O3. A blocking layer was prepared by a spin-coating method using a solution of titanium diisopropoxide bis (acetylacetonate) in 1-butanol with the concentration of 1 M and then heated for 30 min at 500 °C. An electron transfer layer (ETL) prepared by a spin-coating method using TiO2 solution which is obtained by diluting TiO2 (30NR-D) using ethanol (1:6, mass ratio), and then heated for 10 min at 100 °C and 30 min at 450 °C. The UC-Mg-TiO2 was used to fabricate the solar cells by spin-coating a mixed solution of UC-Mg-TiO2 sol and TiO2 sol (UC-Mg-TiO2:TiO2 = x:(100 − x), v/v, x = 0, 20, 40, 60, 80, and 100) on the ETL and heating for 30 min at 500 °C. A perovskite film was fabricated according to the reported method [20]. In brief, the precursor solution of perovskite was prepared by dissolving FAI (1 M), PbI2 (1.1 M), MABr (0.2 M), and PbBr2 (0.22 M) in the mixture of DMF/DMSO (4:1 v:v), and a stock solution of CsI (1.5 M) in DMSO was added. The perovskite film was obtained by spin-coating method with 1000 rpm for 10 s and 4000 rpm for 30 s, and 200 μL chlorobenzene was dropped on the sample before the end of 20 s. A hole transfer layer (HTL) was obtained by the spin-coating method using a spiro-MeOTAD solution at 4000 rpm for 30 s. The spiro-OMeTAD solution was prepared by dissolving 72.3 mg spiro-MeOTAD in 1 mL chlorobenzene and by adding 28.8 μL TBP, 17.5 μL Li-TFSI solution (520 mg/ml in acetonitrile). Finally, an Au anode was made on the hole transfer layer by thermal evaporation.

Characterization

Photoluminescence (PL) spectra were acquired using a fluorometer of FLS 980 E. A diffractometer of DX-2700 was used to obtain the X-ray diffraction (XRD) patterns. X-ray photoelectron spectra were measured with a spectrometer of XPS THS-103. Absorption spectra were obtained with a spectrophotometer of Varian Cary 5000. Scanning electron microscope (SEM) images were performed using a microscope of JSM-7001F. A Keithley 2440 Sourcemeter was applied to measure the photocurrent-voltage (I-V) curves of the solar cells under an illumination of AM 1.5. An electrochemical workstation of CHI660e was utilized to get the electrochemical impedance spectroscopy (EIS). The incident photon-to-current conversion efficiency (IPCE) was measured with a solar cell IPCE recording system (Crowntech Qtest Station 500ADX).

Results and Discussion

The up-conversion fluorescence of the materials was optimized by varying the molar ratio of Ho3+ and Yb3+. The up-conversion emission of Ho3+-Yb3+ co-doped TiO2 with varying molar ratio of Ho3+ and Yb3+ (Ho:Yb:Ti = 1:x:100) was shown in Fig. 1a, which were excited with an 980 nm NIR light. Two strong up-conversion emission peaks were observed at 547 nm and 663 nm. Additional file 1: Figure S1 shows the up-conversion mechanisms of the Ho3+-Yb3+ co-doped TiO2. The fluorescence peaks at 663 nm and 547 nm could correspond to the 5F5 → 5I8 and (5S2, 5F4) → 5I8 transitions of Ho3+, respectively [21]. It can be seen that the intensity of the up-conversion fluorescence is the largest when the molar ratio of Ho3+ and Yb3+ is 1:4. Figure 1b presents the up-conversion photofluorescence of Ho3+-Yb3+-Mg2+ tri-doped TiO2 with different doping contents of Mg2+ (Ho:Yb:Mg:Ti = 1:4:x:100, molar ratio). The up-conversion fluorescence was enhanced by the addition of Mg2+. When the doping content of Ho3+:Yb3+:Mg2+ = 1:4:2, the up-conversion emission is the strongest for Ho3+-Yb3+-Mg2+ tri-doped TiO2. Hereinafter, the UC-Mg-TiO2 with the molar ratio of Ho3+:Yb3+:Mg2+:Ti = 1:4:2:100 was applied.

Fig. 1

Up-conversion emissions of TiO2. a Ho3+-Yb3+ co-doped TiO2 (Ho:Yb:Ti = 1:x:100, molar ratio). b Ho3+-Yb3+-Mg2+ tri-doped TiO2 (Ho:Yb:Mg:Ti = 1:4:x:100, molar ratio)

Figure 2 shows the X-ray diffraction of TiO2 (30NR-D) and UC-Mg-TiO2. According to the PDF card (JCPDS card no.21–1272), the peaks located at 2θ = 25.6 °, 37.7 °, 48.1 °, and 53.7 ° in the patterns could belong to the (101), (004), (200), (105), (211), and (204) crystal planes, respectively. This displays the phase of UC-Mg-TiO2 is anatase.

Fig. 2

X-ray diffraction of TiO2 (30NR-D) and UC-Mg-TiO2

To demonstrate the doping of Ho, Yb, and Mg into TiO2, the X-ray photoelectron spectra of UC-Mg-TiO2 were obtained. The XPS survey spectrum of UC-Mg-TiO2 was presented in Additional file 1: Figure S2. Figure 3a shows the high-resolution photoelectron peaks of Ti 2p, which had two peaks of Ti 2p1/2 and Ti 2p3/2 located at 463.7 eV and 458.2 eV, respectively. Figure 3b, c shows the high-resolution photoelectron peaks of Ho 4d and Yb 4d, which appear at 163.6 eV and 192.3 eV, respectively. These agree with the reported peak positions [22]. Figure 3d presents the photoelectron peak of Mg 2p located at 49.8 eV [23]. These data displays that Ho, Yb, and Mg atoms were incorporated into TiO2.

Fig. 3

X-ray photoelectron spectra of UC-Mg-TiO2. a Ti 2p, b Ho 4d, c Yb 4d, and d Mg 2p

Figure 4a shows the absorption spectra of TiO2 (30NR-D) and UC-Mg-TiO2. There are five absorption peaks appear in the absorption spectrum of UC-Mg-TiO2, which are corresponding to characteristic absorption of Ho3+ and Yb3+. It can be seen that the doping of Ho, Yb, and Mg improves the absorption of TiO2 in visible light region and expands its absorption to NIR range. The Tauc plot can be used to estimate the energy band gap of material [24]. The Tauc plots from the absorption spectra were presented in Fig. 4b. The energy band gap values can be calculated to be 3.09 eV and 3.18 eV for UC-Mg-TiO2 and TiO2 (30NR-D), respectively. The UC-Mg-TiO2 presents a smaller band gap than TiO2.

Fig. 4

a Absorption spectra of TiO2 (30NR-D) and UC-Mg-TiO2. b Tauc plots

Figure 5 shows the SEM photograph of TiO2 (30NR-D) and UC-Mg-TiO2 films. The size of the nanoparticle is about 25 nm for 30 NR-D, and particle size is about 28 nm for UC-Mg-TiO2. The two films are uniform. Thus, the UC-Mg-TiO2 displays a similar morphology and particle size to TiO2 (30NR-D).

Fig. 5

SEM photographs. a TiO2 (30NR-D) film. b UC-Mg-TiO2 film

The PSCs were fabricated based on the electron transfer layers with and without UC-Mg-TiO2. The electron transfer layer with UC-Mg-TiO2 was prepared by spin-coating the mixed solution of UC-Mg-TiO2 sol and TiO2 sol (UC-Mg-TiO2:TiO2 = x:(100 − x), x = 0, 20, 40, 60, 80, and 100, v/v). I-V measurements of the solar cells were performed, and from which the photovoltaic parameters were abstracted. The Isc, Voc, FF, and PCE of the solar cells in this work were obtained by an average of the values of 20 samples. The relation of PCE with the contents of UC-Mg-TiO2 was displayed in Fig. 6a. Firstly, the PCE of the solar cells becomes large, and after that becomes small with the increase of the UC-Mg-TiO2 contents, which reaches the maximum value at the content of 60% (UC-Mg-TiO2:TiO2 = 60:40, v/v). Table 1 presents the photovoltaic parameters of solar cells based on the electron transfer layers with and without UC-Mg-TiO2. The open-circuit voltage (Voc) and short-circuit current (Isc) of the solar cells with UC-Mg-TiO2 were increased to 1.05 V and 22.6 mA/cm2 from 1.03 V and 21.2 mA/cm2 for the solar cells without UC-Mg-TiO2, respectively. Thus, the PCE of the devices based on the electron transfer layer with UC-Mg-TiO2 was improved to 16.3% from 15.2% for those without UC-Mg-TiO2. The typical I-V curves of the devices are shown in Fig. 6b. The PCE histograms of the solar cell performance of 20 samples with and without UC-Mg-TiO2 are presented in Additional file 1: Figure S3.

Fig. 6

a Relationship between the PCE of devices and the contents of UC-Mg-TiO2 (UC-Mg-TiO2 sol: TiO2 sol = x:100 − x, v/v) in the mixed solution. b Typical I-V curves

Table 1

Photovoltaic parameters of the solar cells based on the mesoporous layers with and without UC-Mg-TiO2

Solar cells	Voc (V)	Isc (mA/cm2)	FF (%)	PCE (%)
Without UC-Mg-TiO2	1.03 ± 0.04	21.2 ± 0.7	69.6 ± 1.2	15.2 ± 0.5
With UC-Mg-TiO2	1.05 ± 0.03	22.6 ± 0.6	68.7 ± 1.3	16.3 ± 0.3

Some experiments were carried out to explain the improvement. Figure 7 displays the energy band structures of the materials contained in the solar cells based on some reports [25, 26], and the energy band gap from the Tauc plots is shown in Fig. 4b. The conduction band difference between perovskite and TiO2 becomes larger for UC-Mg-TiO2 compared with that of TiO2 (30NR-D), since the UC-Mg-TiO2 has a smaller band gap than TiO2 (30NR-D). This may be one of the reasons to give a larger Voc for the devices based on the electron transfer layer with UC-Mg-TiO2 [27, 28].

Fig. 7

Energy band structures of the materials contained in the solar cells

Figure 8a shows the steady-state photoluminescence (PL) of the perovskite films on the electron transfer layers with and without UC-Mg-TiO2. The PL peak located at 760 nm is originated from the perovskite film [29]. The PL intensity of the perovskite film on electron transfer layer with UC-Mg-TiO2 decreased compared with that of perovskite film on electron transfer layer without UC-Mg-TiO2. This implies that the electron transport and extraction of UC-Mg-TiO2 from the perovskite film is more efficient than that of TiO2 (30NR-D). This can be further demonstrated by the time-resolved photoluminescence (TRPL) of the samples shown in Fig. 8b. It can be seen that the decay time of TRPL for the perovskite film on electron transfer layer with UC-Mg-TiO2 is faster than that of perovskite film on electron transfer layer without UC-Mg-TiO2. This indicates that the charge transfer for the former is faster than the latter [30, 31].

Fig. 8

a Photoluminescence. b Time-resolved photoluminescence of perovskite film on TiO2 (30NR-D) and UC-Mg-TiO2

Figure 9a shows the Nyquist plots obtained from the electrochemical impedance spectroscopy (EIS) of the solar cells based on the electron transfer layer with and without UC-Mg-TiO2. The Nyquist plots can be fitted by an equivalent circuit which is schematically shown in Fig. 9b. The Rs, Rrec, and Cμ are the series resistance, recombination resistance, and the capacitance of the device [32, 33]. The detailed fitting values are presented in Table 2. The Rs value of the devices based on the electron transfer layers with UC-Mg-TiO2 is nearly the same with that of those without UC-Mg-TiO2. While the Rrec value of the devices based on electron transfer layer with UC-Mg-TiO2 is larger than that of those without UC-Mg-TiO2. This implies that UC-Mg-TiO2 could effectively decrease the change recombination.

Fig. 9

a Nyquist plots obtained from the EIS spectra. b Equivalent circuit utilized to analyze the EIS

Table 2

Fitting parameters for EIS of the devices based on the electron transfer layer with and without UC-Mg-TiO2

Solar cells	Rs/Ω	Rrec/Ω	Cμ-T/F	Cμ-P
With UC-Mg-TiO2	23.4	489.2	9.9E-8	0.8
Without UC-Mg-TiO2	23.1	837.5	13.1E-8	0.8

To confirm the contributions of the up-conversion material UC-Mg-TiO2 to the photocurrent of the solar cells, the I-V measurements were carried out under the simulated solar radiation filtered with a band-pass NIR filter (980 ± 10 nm). Figure 10a displays the I-V curves of the solar cells based on the electron transfer layers with and without UC-Mg-TiO2. The short-circuit current (Isc) of the solar cells with UC-Mg-TiO2 is obviously larger than that of those without UC-Mg-TiO2. This demonstrates the effect of UC-Mg-TiO2 on the photocurrent of the solar cells, because UC-Mg-TiO2 converts the near-infrared photons into visible photons, which the solar cells can absorb to produce additional photocurrent [7, 17]. Figure 10b shows the IPCE spectra of the solar cells with and without UC-Mg-TiO2. The IPCE of the solar cells with UC-Mg-TiO2 is increased, especially at the range of 400~650 nm, compared with that of those without UC-Mg-TiO2. This could be caused by the up-conversion effect of UC-Mg-TiO2 [7, 17].

Fig. 10

a I-V curves of the solar cells under the simulated solar radiation filtered with a band-pass NIR filter (980 ± 10 nm). b IPCE spectra of the solar cells with and without UC-Mg-TiO2

Conclusions

The up-conversion nanomaterial of Ho3+-Yb3+-Mg2+ tri-doped TiO2 (UC-Mg-TiO2) was synthesized successfully. The up-conversion emissions of the UC-Mg-TiO2 were enhanced with an addition of Mg2+. We applied the UC-Mg-TiO2 to the PSCs, in which the UC-Mg-TiO2 was used to modify the electron transfer layer. The Voc and Isc of the devices with UC-Mg-TiO2 were improved to 1.05 V and 22.6 mA/cm2 from 1.03 V and 21.2 mA/cm2 for those without UC-Mg-TiO2, respectively. And the PCE of the devices with UC-Mg-TiO2 was increased to 16.3% from 15.2% for those without UC-Mg-TiO2.

Figure S1. Up-conversion mechanisms of the Ho3+-Yb3+ co-doped TiO2. Figure S2. XPS survey of UC TiO2. Figure S3 PCE histograms of the solar cell performance of 20 samples with and without UC-Mg-TiO2. (DOCX 77 kb)