1. Introduction

In order to solve the problem of the long-term worldwide energy demand, the photovoltaic cells which can convert sunlight into electricity have become more and more important [1]. Among them, the crystalline Si (c-Si) solar cells occupy a majority of the solar cells market. However, the low energy efficiencies around 15% is the chief obstacles of the crystalline Si (c-Si) solar cells for its mass-application [2]. The mismatch between the solar spectrum and the band gap energy of silicon semiconductor is the main reason for source of energy loss (over 70%) [2, 3]. Due to the lower absorption rate of sunlight, the photoelectric conversion efficiency of solar cells is low. Hence, how to increase the photoelectric energy conversion efficiency of solar cells is very meaningful. To increase the energy efficiency, two general approaches were proposed. One is to develop tandem solar cells consisted of multiple semiconductor layers and each semiconductor layer has a characteristic band gap converting a different part of the solar spectrum. However, the complicated fabrication and high cost limit their practical application [3, 5]. The other one has been achieved by quantum cutting (QC). QC is based on the process that an incident high-energy photon splits into two or more low-energy photons, which can reduce the energy loss related to the thermalization of hot charge carries after the absorption of a high-energy photon. One ultraviolet (UV) or blue photon can be cutted to two near-infrared photons which can both be absorbed by solar cell [4–7]. In this case, the energy loss can be effectively reduced and the near-infrared quantum cutting (NIR-QC) has attracted a lot of research attention. Most notable is the luminescent solar concentrator (LSC) based on the theory of NIR-QC. It has become a unique and efficient photovoltaic optical device which consists of a luminescent material dispersed in a transparent waveguide plate as well as Si solar cells optically matched to the plate edges [2, 6, 7]. It is well known that the QC luminescent materials can modify spectral efficiently by combination of different rare earth ions thanks to their abundant energy levels. NIR-QC was first achieved in YPO₄: Tb³⁺, Yb³⁺ [8]. In this system, one blue photon of Tb³⁺ ion from the absorption transition of ⁷F₆-⁵D₄ around 483 nm is cut into two NIR photons of Yb³⁺ ion by the emission of ²F_5/2-²F_7/2. The cooperative energy transfer (CET) from the donor (Tb³⁺) to the acceptor (Yb³⁺) via downconversion is involved with the NIR-QC process. Therefore, it is expected that NIR-QC luminescent materials would significantly maximize LSC performance with Si solar cells and other application.

So far, similar NIR-QC phenomena with RE/Yb³⁺ (RE = Tb³⁺, Pr³⁺, Tm³⁺, Er³⁺, Mn²⁺, Eu²⁺) couple were reported in various powder and glass systems [9–14]. Due to the Yb³⁺ can emit photons around 1000 nm [25, 26], it often introduced as a sensitizer in these materials. Nevertheless, the NIR luminescence of Yb³⁺ duo to CET process is still weak. It is mainly ascribed to the inefficient excitation of the donors for their parity forbidden 4f-4f transitions, i.e., low absorption cross sections, typically on the order of 10⁻²¹ cm². The absorption of Ce³⁺ is originated from the allowed electric-dipole transition from the 4f ground state to the 5d excited one, which results in a very high absorption cross section in an order of 10⁻¹⁸ cm² in the ultraviolet (UV) region [6, 23, 24]. Moreover, the excitation band of Ce³⁺ is broad and the excitation wavelength of Ce³⁺ can be tuned by the host materials, which means that high quantum efficiency (QE) may be obtained by choosing appropriate host materials. All these conditions show that strong NIR emission with high QE may be obtained in Ce³⁺, Yb³⁺ co-doped materials. The QC of visible light to NIR emission in Ce³⁺, Yb³⁺ co-doped Y₃Al₅O₁₂ [15], YBO₃ [16], LuBO₃ [17] and glasses phosphors [18] had been reported recently. In these systems, a high QE will be possible at relatively high Yb³⁺ concentration where the NIR emission is quenching at a low concentration of Yb³⁺ (<10%). In this paper, we develop a novel NIR-QC KCaGd(PO₄)₂:Ce³⁺, Yb³⁺ phosphor. It can harvest UV photons and emit NIR photons around 1021 nm efficiently. The NIR emission intensity of KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ didn’t quenched at 20% Yb³⁺ doping and the highest NIR QE is about 158.2%. KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ is a promising solar spectral converter and concentrator for LSC with Si solar cells.

2. Experiments

Rare earth ion doped samples of KCaGd(PO₄)₂ were prepared by the thermal decomposition of the corresponding nitrate. The starting materials were CaCO₃ [analytical reagent (A. R.)], K₂CO₃ (A. R.), Gd₂O₃ (99.99%), Yb₂O₃ (99.99%), CeO₂ (99.99%) and (NH₄)₂HPO₄ (98.5%). Three steps were necessary for synthesizing the samples. Firstly, stoichiometric amounts of CaCO₃, K₂CO₃, (NH₄)₂HPO₄ and rare earth oxides were thoroughly mixed, dissolved completely in the nitric acid, and a certain amounts of deionized water was added into this solution until it is transparent. Secondly, the transparent solution was heated to evaporate water and excess HNO₃, at the same time, the coprecipitation was obtained and then dried at 80°C for 10h. Finally, the resulting mixture is heated at 900°C in a reductive atmosphere (H₂ + N₂) for 6 h and cooled to the room temperature. The obtained samples were white powder bulk.

The X-ray powder diffraction (XRD) patterns were obtained on Rigaku D/max-2000 powder diffract meter using Cu Kα radiation (1.5405 Å) at 40 kV and 60 mA. The absorption spectra were measured with a visible-NIR spectrophotometer (PE Lambda 950) using BaSO₄ as a reference. Emission and excitation measurements were performed using a 450 W Xe lamp as the excitation source. Mean decay lifetime measurements were performed using uF900 as the excitation source. All measurements were carried out at room temperature.

3. Results and discussion

Figure 1 shows the XRD patterns of KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺(0≤x≤0.2) samples and the reference data of JCPDS Card No. 34-0125 for pure KCaGd(PO₄)₂. It was reported that KCaGd(PO₄)₂ had a hexagonal crystal structure with the space group P6₂22 [19, 20]. Compared with standard data, all the samples exhibit the peaks of pure hexagonal phase. No second phase is detected in the XRD patterns, revealing the successful doping of Ce³⁺ and Yb³⁺ ions in KCaGd(PO₄)₂. To obtain a detailed crystal structure regarding the host KCaGd(PO₄)₂, Rietveld refinement was performanced with the crystallographic data taken from the structure of LaPO₄ (ISCD NO.33241).

 

Fig. 1 Powder XRD patterns of KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺(0≤x≤0.2).

Download Full Size | PDF

Figure 2 shows the observed (crosses), calculated (solid line), and difference (bottom) XRD profiles for the Rietveld refinment of KCaGd_0.98Ce_0.01Yb_0.01(PO₄)₂. The Rietveld analysis results indicate that the weighted profile R-factor(R_wp) and the excepted R factor (R_e) are 13.37% and 10.35%, respectively. The results of the final refinement data indicate that the powder sample is well crystallized with space group P6₂22 and lattice parameters a = b = 6.9521 Å, c = 6.3680Å. KCaGd(PO₄)₂ has the similar crystal structure with hexagonal LaPO₄, except that the large 8-coordinated K⁺ occupies only half of the tunnel sites with the coordination polyhedron [KO₈], and La³⁺ position is statistically occupied by Gd³⁺ or Ca²⁺ ions [19, 20]. As shown in the inset of Fig. 2, the unit cell parameters a and c decrease linearly with the increasing Yb³⁺ concentration, which accords with the Vegard’s rule [27]. The linear decreasing trend of unit cell parameters is due to the substitution of Gd³⁺ ions by small Yb³⁺ ions in the host lattice. All these results indicate that the hexagonal phase KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ is fully developed and a small amount of Ce³⁺ and Yb³⁺ ions are successfully doped into the host.

 

Fig. 2 Rietveld refinement of the powder XRD pattern of KCaGd_0.98Ce_0.01Yb_0.01(PO₄)₂ (observed—cross, calculated—black line, difference between the observed and the calculated—bottom blue line, and Bragg positions—vertical bars). Inset shows unit cell parameters a (A°), and c (A°) dependence on Yb3 + concentration in KCaGd(PO4)₂: 0.01Ce³⁺, xYb³⁺(0≤x≤0.2).

Download Full Size | PDF

The absorption spectra of KCaGd(PO₄)₂: 0.01Ce³⁺, KCaGd(PO₄)₂: 0.0.1Ce³⁺, 0.2Yb³⁺ and BaSO₄ are shown in Fig. 3. The thickness of sample KCaGd(PO₄)₂: Ce³⁺ and KCaGd(PO₄)₂: Ce³⁺,Yb³⁺ was around 1mm. In the range of 850 nm-1100 nm, the absorption intensity of KCaGd(PO₄)₂: 0.01Ce³⁺, 0.2Yb³⁺ is increased compared with Yb³⁺ undoped sample and BaSO₄, so that the characteristic absorption peaks due to Yb³⁺ can be observed more clearly. And the strong absorption band assigned to the Yb³⁺: ²F_7/2-²F_5/2 transition is centered at 980 nm in the Yb³⁺ doped samples [21]. Other strong absorption band of KCaGd(PO₄)₂: 0.01Ce³⁺ and KCaGd(PO₄)₂: 0.0.1Ce³⁺, 0.2Yb³⁺ samples assigned to the 4f-5d transition of Ce³⁺ is also visible around 315 nm.

 

Fig. 3 Absorption spectra of KCaGd(PO₄)₂: 0.01Ce³⁺, KCaGd(PO₄)₂: 0.01Ce³⁺, 0.2Yb³⁺ and BaSO₄.

Download Full Size | PDF

In order to give a convincing evidence for the energy transfer of Ce³⁺ to Yb³⁺, luminescence measurements of Ce³⁺ and Yb³⁺ and decay curves of Ce³⁺ in KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ phosphors are investigated. Figure 4 presents the photoluminescence (PL) and photoluminescence excitation (PLE) spectra of the KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺ (0≤x≤0.2) samples . By monitoring the 5d to 4f transition of Ce³⁺ at 366 nm, two broad bands were observed from 260 to 350 nm with a maximum at about 280 nm and 324 nm, respectively. These two broad bands can be attributed to the f-d transitions of Ce³⁺ in the host lattice, when Ce³⁺ enters one specific site, its 5d state will be split into several different components depending on the site symmetry [22]. Under 324 nm excitation, the emission spectra shows the d-f transition band in the region of 340-450 nm. It is noticed that, as the Yb³⁺ concentration increases from 0% to 20%, the intensity of d-f transition gradually decreased. This indicates that an efficient energy transfer from 5d energy level of Ce³⁺ to ²F_5/2 energy leve of Yb³⁺ ions could occur.

 

Fig. 4 (a) Excitation (λ_em = 366nm) and (b) emission (λ_ex = 324nm) spectra of Ce³⁺ in KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺

Download Full Size | PDF

To verify this energy transfer process, the excitation spectra monitored by 1021 nm emission wavelength are depicted in Fig. 5(a). As shown in the Fig. 5(a), all the excitation spectra with different Yb³⁺ content are similar to the KCaGd(PO₄)₂: 0.01Ce³⁺, 0.1Yb³⁺ sample except the intensity. The broad excitation band in the region of 240-350 nm owing to the d-f transition of Ce³⁺ is similar with Fig. 4(a). This result further confirms the existence of energy transfer from Ce³⁺ to Yb³⁺. As we known, the energy part bellow 400 nm of the AM1.5 spectrum is only 3% of the solar photons arriving on the surface of our planet. Due to the lower absorptivity of the sunlight, the KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ phosphors also have deficiencies in the practical application. However it also can be considered to be the strength of the material compared to other materials which can absorb the visible light. Such as Tb³⁺-Yb³⁺/Pr³⁺-Yb³⁺/Tm³⁺-Yb³⁺ co-doped phosphors and so on, these materials can absorb the visible light in the range of 440-490 nm. In this materials, due to the inefficient excitation of the donors, the NIR luminescence of Yb³⁺ duo to energy transfer process is still weak. But the absorption of Ce³⁺ is originated from the allowed electric-dipole transition from the 4f ground state to the 5d excited one, which results in a very high absorption cross section in the UV region and caused to the efficient energy transfer from Ce³⁺ to Yb³⁺. Figure 5(b) shows the NIR emission spectra of KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺ under 324 nm excitation. It is worthy to note that a clear emission centered at 1021 nm was also observed, corresponding to the ²F_5/2-²F_7/2 transition of Yb³⁺, and the intensity of the Yb³⁺ emission enhances with the increase of the Yb³⁺ concentration. The concentration quenching has not been discovered until the Yb³⁺ content increased to 20%. To add the proof for the energy transfer from Ce³⁺ to Yb³⁺, the Yb³⁺ single doped KCaGd(PO₄)₂ sample are also excited by 324 nm as reference. It can be seen that the emission of Yb³⁺ from transition of ²F_5/2-²F_7/2 cannot be observed in the Yb³⁺ single doped KCaGd(PO₄)₂ sample obviously, as might be expected. Thus, we can identify that the efficient CET from Ce³⁺ to Yb³⁺ is existent in this system. In this work, we investigated the quantum cutting in KCaGd(PO₄)₂: Ce³⁺, Yb³⁺, whose theory is based on the quantum cutting model of cooperative energy transfer that has been reported in the references [8–11, 15–18]. Therefore, the emission of two NIR photons per absorbed one UV photon is possible with the Ce³⁺-Yb³⁺ dual ions via this CET process.

 

Fig. 5 (a) Excitation (λ_em = 1021 nm) and (b) near-infrared emission (λ_ex = 324 nm) spectra of Yb³⁺ in KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺ and KCaGd_0.99-x(PO₄)₂: 0.01Yb³⁺.

Download Full Size | PDF

Figure 6 illustrates the schematic energy levels with transitions which may be involved in the CET process from one Ce³⁺ to two Yb³⁺. Upon excitation at 324 nm, Ce³⁺ emissions occur at 366 nm (shown in Fig. 4), which can be assigned to the transition from the lowest 5d level to ²F_5/2 and ²F_7/2 of Ce³⁺, respectively. Just as mentioned in the reference 8-11 and 15-18, different donor transfer the energy to two acceptor (Yb³⁺) from their higt energy level by cooperative energy transfer, because the energy of the donor’s high level is twice as the excited state of acceptor. In the KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ phosphor, the energy of Ce³⁺ 5d level (around 3000 cm⁻¹) is almost triple as ²F_5/2 excited state of Yb³⁺. Considering the cooperative energy transfer theory mentioned above, we can conclude that two NIR photons (around 1021 nm) are obtained due to the ²F_5/2→²F_7/2 transition of Yb³⁺. Under this circumstance, the CET of Ce³⁺: 5d→Yb³⁺: ²F_5/2 + Yb³⁺: ²F_5/2 is the only possible relaxation route to achieve the Yb³⁺ NIR emission because the Ce³⁺: 5d→4f emission is located at approximately twice the energy of the Yb³⁺: ²F_5/2→²F_7/2 absorption and Yb³⁺ has no other levels up to the UV region. Ce³⁺, Yb³⁺ co-doped KCaGd(PO₄)₂ phosphors present two attractive features: firstly, the conversion of one UV photon in the region of 200–350 nm (almost useless in the silicon solar cell) to two NIR photons around 1021 nm (where the silicon solar cell exhibits the greatest spectral response). And secondly, the highly efficient Ce³⁺ excitation and Yb³⁺ NIR luminescence benefit from the large absorption cross section of Ce³⁺ [15–18]. In the cooperative energy transfer mechanism, the transition from lower excited state to higher virtual state is inefficient. But the resonant energy transfer is an efficient mechanism for rare-earth ions due to the energy directly transfer from higher energy level of donor to lower energy level of acceptor by means of resonance. Hence the efficiency of energy transfer is lower, because the cooperative energy transfer is clearly off resonance.

 

Fig. 6 Schematic energy level diagram and cooperative energy transfer mechanism of Ce³⁺ and Yb³⁺ in KCaGd(PO₄)₂: Ce³⁺, Yb³⁺.

Download Full Size | PDF

Figure 7(a) shows the Yb³⁺ concentration dependence of the luminescence decay curves and (b) fitted results of Ce³⁺ emission for all the powder samples KCaGd_0.99-x(PO₄)₂: 0.01Ce³⁺, xYb³⁺ (0≤x≤0.2) under 324 nm excitation at the room temperature. The Ce³⁺: 5d→4f emission in the Ce³⁺ single doped sample shows, as expected, a nearly single exponential decay, and the fitting yields a decay time of 25.1 ns. The lifetime in nanosecond order is one of the characteristics of the Ce³⁺ electric-dipole allowed 5d→4f transition [18]. In the whole Yb³⁺ doping concentration range 0≤x≤0.2, Ce³⁺ decays single exponentially, and its lifetime is deceased from 25.1 ns when no Yb³⁺ is codoped to 10.5 ns when Yb³⁺ concentration is increased to 20%. Under direct Ce³⁺ f-d transitions at 324 nm, the decrease of Ce³⁺ d-f emission intensity along with the decrease of Ce³⁺ decay time, providing a clear evidence that Ce³⁺→Yb³⁺ energy transfer takes place in this system. From the luminescence decay curves in Fig. 7 the energy transfer efficiency and the quantum efficiency can be determined. The energy transfer efficiency η_ETE is defined as the ratio of Ce³⁺ that depopulate by energy transfer to Yb³⁺ ions over the total number of Ce³⁺ ions excited. We define the total quantum efficiency, η_QE, as the ratio of the number of photons emitted (visible and infrared) to the number of photons absorbed, assuming that all excited Yb³⁺ decay radiatively [28]. By dividing the integrated intensity of the decay curves of the Ce³⁺, Yb³⁺ codoped samples to the Ce³⁺ single-doped sample, η_ETE is obtained as a function of the Yb³⁺ concentration as follows: [8,10,]

 

Fig. 7 Luminescent decay curves (a) and the fitted decay curves (b) of the 366 nm emission of Ce³⁺ in KCaGd(PO₄)₂: Ce³⁺, Yb³⁺ samples with different Yb³⁺ concentration (λ_ex = 324 nm). The scale of y-axis is semi-log.

Download Full Size | PDF

Table 1. The decay lifetime of Ce³⁺, the energy transfer efficiency η_ETE and the quantum efficiency η_QE as a function of Yb³⁺ doping concentration in KCaGd(PO₄)₂: Ce³⁺, Yb³⁺.

View Table

In Fig. 8, decay lifetime and η_QE are plotted as a function of Yb³⁺ concentration. It is presented that the decay lifetime decreases from 25.1 ns to 10.5 ns with the increase of Yb³⁺ content from 0 to 20%, and the quantum efficiency increases from 0 to 158.2%. Supposing there is no loss of nonradiation by defects and impurities, the maximum quantum efficiency of 158.2% is obtained.

 

Fig. 8 Decay lifetime of Ce³⁺ and quantum efficiency plotted as a function of Yb³⁺ concentration.

Download Full Size | PDF

4. Conclusions

We have successfully developed a novel NIR QC KCaGd(PO₄)₂:Ce³⁺, Yb³⁺ phosphor. It has intense wide excitation bands in the UV region and exhibits a greatly enhanced NIR emission of Yb³⁺ around 1021 nm through CET from one Ce³⁺ to two Yb³⁺, perfectly matching the maximum spectral response of Si solar cells. The theoretical highest QE has been found to be 158.2% with the doping concentration 20% of Yb³⁺. These results indicate that KCaGd(PO₄)₂:Ce³⁺, Yb³⁺ phosphor could improve the efficiency of silicon-based solar cells by means of downconversion.