Introduction

Solar energy has been considered as a green and renewable alternative energy source to traditional fossil fuels1. Since the prototype of a dye-sensitized solar cell (DSSC) was reported in 1991 by Grätzel2, it has aroused intensive interest and become one of hotspots in solar energy field because of its ease of fabrication and cost-effectiveness compared with silicon-based photovoltaic devices3,4. Recently, some important progresses are achieved5,6,7,8,9,10. However, the efficiencies of the DSSCs are lower than that of Si solar cells, which restrict DSSC's potential application. An effective method for enhancing the efficiency is broadening the absorption range of the DSSC. Consequently, many metal complexes dyes have been synthesized. But even the best of these (N-719, N-749, YD2-o-C8) only absorb visible light in the wavelength range of 400–800 nm3,4 and most of the solar ultraviolet and infrared irradiations are not utilized. Recently, the researches on the energy relay dyes (ERDs) via Forster resonant to broaden the absorption domain and thus increase the photocurrent have been done11,12. Another alternative route for widening absorption range is the conversion luminescence by doping rare-earth compounds. On the other hand, rare-earth ions are +3 value cations, when they are doped into TiO2 semiconductor, a p-type doping effect occurs13,14, which results in the elevation of Fermi level of the photoanode and turns to the enhancement of the photovoltage of the DSSC. The introduction of doped rare-earth compound not only increases the photocurrent via conversion luminescence, but also improves the photovoltage by p-type doping effect, this double functions is very significant for enhancing the photovoltaic performance of the DSSC. Unfortunately, little significant research on conversion luminescence and/or p-type doping effect by rare-earth ions in the DSSC has been attempted15,16,17.

Results

Phase analysis of YF3:Eu3+

The YF3:Eu3+ was prepared by hydrothermal method18. The X-ray diffraction (XRD) pattern of prepared YF3:Eu3+ is shown in Fig. 1. All diffraction peaks of the prepared sample are readily indexed as orthorhombic phase of YF3 and are consistent with the standard pattern (JCPDS 74-0911). This indicates the formation of the orthorhombic phase of YF3. Furthermore, no EuF3 phases are observed, this is because that in the preparation Y2O3 and Eu2O3 are completely dissolved and mixed (see method section), Eu3+ amount in YF3 mixed solution is small (2.0 mol.%) as well as Y3+ and Eu3+ ions have similar electronic structures (s2p6) and radii (0.90Å and 0. 95Å), therefore, Eu3+ ions occupy the lattice sites of Y3+ ions.

Figure 1
figure 1

XRD patterns of YF3:Eu3+.

Full size image

Photoluminescence properties of YF3:Eu3+

The excitation spectrum of YF3:Eu3+ (emission at 592 nm) is shown in Fig. 2a. The excitation spectrum of YF3:Eu3+ consists of a number of line peaks in the range of 318–475 nm and a main peak appears at 393 nm. The typical line peaks located at 318, 361, 383, 393 and 464 nm are related to the 4f electronic transitions of Eu3+ ions: 7F0 → 5H6, 7F0 → 5D4, 7F0 → 5G2, 7F0 → 5L6 and 7F0 → 5D2, respectively, which are accorded with the literatures18,19,20. These results indicate that the Eu3+ ions can be excited with ultraviolet light either indirectly through the YF3 host lattice or directly through absorption by Eu3+ ions themselves.

Figure 2
figure 2

Excitation (a) and emission (b) spectra of YF3:Eu3+.

Full size image

Fig. 2b shows the emission spectrum of YF3:Eu3+ under 393 nm excitation, consisting of line peaks mainly locate at 592, 612, 650 and 701 nm, corresponding to the transitions of Eu3+ ions: 5D0 → 7FJ (J = 1, 2, 3, 4)21,22,23. These luminescence bands are located in the absorption range of the N-719 dye. Combining the excitation and emission spectra, the ultraviolet irradiation can be absorbed by the N-719 dye in the DSSC via down-conversion luminescence, which widens the light absorption range of the DSSC.

Electrochemical analysis for YF3:Eu3+ doped TiO2 film

The influence of YF3:Eu3+ on the energy level of TiO2 is assessed through Mott-Schottky electrochemical analysis24,25. Fig. 3a shows Mott-Schottky plots (1/C2 vs. V) of the TiO2 film with different amount of YF3:Eu3+, from which the flat-band potential (VFB) can be obtained (Table 1). It can observe a negative shift in the flat-band potential with the increase of YF3:Eu3+ percentage from the −0.44 V for pure TiO2 film to −0.83 V for the film doped with YF3:Eu3+ of 7 wt%. According to the doping principle, pristine TiO2 is a n-type semiconductor13,14, the introduction of +3 value metal ions into +4 value metal oxide (TiO2) will cause a p-type doping effect (similar to that +3 value ions doped in Si semiconductor). Eu3+ ion is +3 value rare-earth ion, so Eu3+ ion doping is a p-type doping, which leads to the negative shift in the flat-band potential with the increase of YF3:Eu3+ percentage in the TiO2 film. It is notable that the doping does not change TiO2 semiconductor type, but it changes the Fermi level and the flat-band potential of the TiO2.

Table 1 Flat-band potential (VFB) and saturated current density (ISAT) for YF3:Eu3+/TiO2 systems
Full size table
Figure 3
figure 3

Mott-Schottky plots (a) and current-potential curves (b) of YF3:Eu3+/TiO2 systems.

Full size image

The potentiodynamic parameters of the YF3:Eu3+/TiO2 films were measured by electrochemical method24,25. Fig. 3b shows the current-potential curves of YF3:Eu3+/TiO2. For all samples, with the increase of the applied potential, the photocurrent densities firstly increase and then gradually tend to a saturation value. On the other hand, the photocurrent densities for the YF3:Eu3+/TiO2 are all larger than that for TiO2 alone. As shown in Fig. 3b and Table 1, the saturated current density for TiO2 alone is about 10.12 μA·cm−2 and the saturated photocurrent densities for all YF3:Eu3+/TiO2 are larger than this.

Photovoltaic performance of DSSCs

To study the effect of ultraviolet irradiation on the DSSC, we used a filter stop (ZWB3) to remove the light with wavelength larger than 420 nm. Fig. 4a shows the photovoltaic character curves of the DSSC with and without YF3:Eu3+ doing under an ultraviolet irradiation of 24 mW·cm−2. The DSSC without YF3:Eu3+ shows the following photovoltaic parameters: short circuit current density (JSC) = 1.192 mA·cm−2, open circuit voltage (VOC) = 0.584 V, fill factor (FF) = 0.536 and light-to-electric energy conversion efficiency (η) = 0.373%. However, the DSSC containing YF3:Eu3+ displays enhanced photovoltaic parameters: JSC = 1.406 mA·cm−2, VOC = 0.691 V, FF = 0.573 and η = 0.557%. This enhancement indicates that ultraviolet light can be effectively converted to visible light by YF3:Eu3+ and the more incident light are harvested, therefore, the efficiency of the DSSC is improved.

Figure 4
figure 4

J–V curves of DSSC with and without YF3:Eu3+ doping (5 wt.%) under an ultraviolet light irradiation of 24 mW·cm−2 (a) and J–V curves of the DSSC with different amount of YF3:Eu3+ doping under a simulated solar light irradiation of 100 mW cm−2 (b).

Full size image

The photocurrent-voltage curves of the DSSCs with and without YF3:Eu3+ doping under a simulated solar light irradiation of 100 mW·cm−2 were measured and shown in Fig. 4b and the photovoltaic parameters are listed in Table 2. Obviously, with the increase of YF3:Eu3+ amount in the DSSC, VOC values increase, JSC values first increase and then decrease. When the amount of YF3:Eu3+ is 5 wt.% in the doping layer, the DSSC achieves a maximum efficiency of 7.741%. The increase in JSC when the doping amount is less than 5 wt.% can be attributed to the down-conversion luminescence of Eu3+ from ultraviolet light to visible light. Moreover, the changes of TiO2 energy level by doping rare-earth ions improve the carrier transport at the interface of TiO2/Dye, which further improves the JSC. However, when the amount of YF3:Eu3+ is beyond 5 wt.% in the doping layer, more grain, phase and domain interfaces are produced in the doping layer. These interfaces can capture photogenerated electrons and holes, hinder the charge carrier transportation, leading to a decrease in photocurrent26,27. This is why JSC increases and then decreases with the increment of YF3:Eu3+ amount in the doping layer of the DSSC.

Table 2 Effect of YF3:Eu3+ on photovoltaic properties of the DSSCs
Full size table

On the other hand, as shown in Fig. 4b and Table 2, the VOC values increase with the increase of YF3:Eu3+ contents. When the amount of YF3:Eu3+ is 7 wt% in the doping layer, the DSSC obtains a VOC value of 0.800 V, increasing 10% compared to the DSSC without YF3:Eu3+ doping. According to DSSC operation principle3,4, the VOC corresponds to the energy difference between the electronic Fermi level of TiO2 film and the redox potential of the electrolyte (VOC = EF(TiO2) − Eredox). When Eu3+ or Y3+ ions are doped and substituted for the Ti4+ ion lattice sites in TiO2, it will give a p-type doping effect, similar to that M3+ ions doped in Si crystal for Si solar cells, which results in the elevation of the Fermi level of TiO2 electrode (more negative vs vacuum) and the increase of VOC values28.

As shown in Table 2, when the ratio of YF3:Eu3+/TiO2 is 5 wt.% in the doping layer, the light-to-electric energy conversion efficiency of the DSSC reaches 7.741%, increasing by 32% compared to the DSSC without the rare-earth ions doping. The enhanced degree of photovoltaic performance of the DSSC by doping rare-earth fluoride is higher than that by doping rare-earth oxide15,16,17.

Discussion

Different from other researches, in which spectral converters by rare-earth compound were installed on the front or the back29, we introduced doped rare-earth compound (YF3:Eu3+) in the TiO2 photoanode (doping layer) of DSSC, which resulted in more effective conversion luminescence and p-type doping effect. As a conversion luminescence medium, YF3:Eu3+ broadens ultraviolet light harvest via down-conversion luminescence and increases photocurrent; As a p-type dopant, Eu3+ ions elevate the Fermi level of the TiO2 film and thus heightens the photovoltage. When the doping amount is 5 wt.% in the doping layer, the light-to-electric energy conversion efficiency of the DSSC reaches 7.74%, which is increased by 32% compared to that of the DSSC without YF3:Eu3+. The present results demonstrate the feasibility of the conversion luminescence and p-type doping effect by doped rare-earth compound in the DSSC and provide a novel route for enhancing the photovoltaic performance of solar cells.

Methods

Materials

Chemical reagents including tetrabutyl titanate, polyethylene glycol (molecular weight of 20000), 4-tert-butylpyridine (TBP), oxalic acid, nitric acid, OP emulsifying agent (Triton X-100), iodine, lithium iodide, tetramethyl ammonium iodine, acetonitrile, isopropanol, yttrium oxide, europium oxide, hydrochloric acid, hydrofluoric acid, ammonium hydrogen fluoride are analytic purity from Shanghai Chemical Agent Ltd., Shanghai, China. The sensitized dye N-719 [RuL2(NCS)2, L = 4,4′-dicarboxylate-2,2′-bipyridine] was from Solaronix SA, (Aubonne, Switzerland).

Preparation of YF3:Eu3+

YF3:Eu3+ powder was prepared by hydrothermal method according to the following procedures18. Firstly, Y2O3 (0.100 mol) and Eu2O3 (0.002 mol) were mixed and homogenized thoroughly. Then, hydrochloric acid was added into the mixture under heating and stirring. After being completely dissolved, the mixture was quickly transferred into a Teflon-lined stainless-steel autoclave, then, NH4HF2 (0.500 mol) was added into the autoclave under stirring. Afterwards, appropriate de-ionized water was added until the filled degree reached 75–80% of the total container volume. Then, pH value of the mixed solution was adjusted to 4.5 using hydrofluoric acid. The obtained solution was hydrothermally treated at 200°C for 4 days. After being naturally cooled to room temperature, the obtained product was centrifuged, washed until the pH value of the system equalled to 7 and then dried in air at ambient temperature. Thus, YF3:Eu3+ powder was obtained.

Preparation of film electrodes and fabrication of DSSC

The TiO2 colloid was prepared as reported previously17,30,31. Then, the YF3:Eu3+ powder was dispersed into the TiO2 colloid by ultrasonically vibrating for 90 min and hydrothermally treating at 200°C for 24 h to form a (TiO2 + YF3:Eu3+) colloid. A TiO2 film with a thickness of about 12 μm was prepared by coating the TiO2 colloid on the FTO plate. Then, sintered at 450°C for 30 min. Afterwards, the doping layer of (TiO2 + YF3:Eu3+) with a thickness of about 4 μm was coated on the TiO2 film in the same way. Then, the film was soaked in dye N-719 for 24 h to form a dye-sensitized film electrode. For comparison, a dye-sensitized TiO2 film electrode without YF3:Eu3+ was made, the film thicknesses was about 16 μm, which was the same as the film with YF3:Eu3+ doping.

The DSSC was assembled by injecting the electrolyte (0.6 M tetra butyl ammonium iodide, 0.1 M iodine, 0.1 M lithium iodide, 0.5 M 4-tert-butyl-pyridine (TBP) in acetonitrile) into the aperture between the dye-sensitized TiO2 film electrode and a platinum counter electrode. The two electrodes were clipped together and a cyanoacrylate adhesive was used as sealant to prevent the electrolyte solution from leaking17,30,31.

Measurement and characterization

The phase purity and crystal structure of YF3:Eu3+ powders were analyzed with the powder X-ray diffraction (XRD) measurements. The XRD patterns were recorded by an X-ray diffractometer (BRUKERD8, Karlsruhr, Germany) using Cu-Kα radiation (λ = 1.5405 Å). The 2θ angle of the XRD spectra was recorded at a scanning rate of 3° min−1.

The excitation and emission spectra were measured using a spectrophotometer (FLS920, Edinburgh, UK), in which a xenon lamp and a photomultiplier tube (R955, Hamamatsu) were equipped as excitation source and the detector of fluorescence, respectively.

Electrochemical experiments were conducted on an electrochemical cell filled with an electrolyte solution contained 0.01 M LiClO4, 10 mM LiI and 1 mM I2 in acetonitrile solvent, where a Pt counter electrode, a saturated calomel reference electrode (SCE) and a working electrode were immersed. The capacitance-voltage (C-V) curves24,25 and potentiodynamic current-applied potential (I-V) curves32,33 were measured by using an Electrochemical Workstation (CHI660C, Shanghai Chenhua Device Company, China). The data of C-V were used for evaluating the flat band potential according to Mott-Schottky equation (Eq. 1)24,25, which relates the capacity to the potential difference between the surface and the bulk of a semiconductor ΔV:

In the above equation, ε is the dielectric constant of the material, ε0 is the permittivity of the vacuum, e is the element charge, Ni is the concentration of donors in n-type and acceptors in p-type semiconductor, Zi is +1 for donors and −1 for acceptors.

The photovoltaic parameters of the DSSC were characterized by measuring the J–V character curves under a simulated solar light irradiation of 100 mW·cm−2 from a 100 W Xe lamp (XQ-500 W, Shanghai Photoelectricity Device Company, Shanghai City, China). The fill factor (FF) and light-to-electric energy conversion efficiency (η) of the cell were calculated according to the following equations3:

where JSC is the short-circuit current density (mA·cm−2), VOC is the open-circuit voltage (V), Pin is the incident light power and Jmax (mA·cm−2) and Vmax (V) are the current density and voltage in the J–V curves at the point of maximum power output, respectively.