Yan Chen*a,
Jing Wang*b,
Mei Zhanga and
Qingguang Zenga
aSchool of Applied Physics and Materials, Wuyi University, Jiangmen, Guangdong 529020, P. R. China. E-mail: ychen08@163.com; Fax: +86 20 84112112
bSchool of Chemistry, Sun Yat-sen University, Guangzhou, Guangdong 510275, P. R. China. E-mail: ceswj@mail.sysu.edu.cn
First published on 18th April 2017
A light conversion material LiBaPO4:Eu2+, Pr3+ is successfully developed as a solar spectral converter for Si solar cells. The photoluminescence excitation and emission spectra from 3 K to room temperature, the lifetime and energy transfer mechanism are all systematically investigated. The results show that Eu2+ ions absorb and transfer the energy of UV-Vis photons to Pr3+ ions, which exhibit an intense near-infrared (NIR) emission between 950–1060 nm, matching well with the maximum spectral response of Si solar cells. We believe that this NIR emitting phosphor may open a new route to design advanced UV-Vis-to-NIR phosphors, with promising applications for solar spectral convertors.
Recently, light conversion material is used to modify the solar spectrum and enhance the external quantum efficiency of Si solar cell.6–8 Many researches have been developed phosphors doped by Yb3+ ion, Eu3+–Yb3+, Tb3+–Yb3+, Dy3+–Yb3+, Eu2+–Yb3+ and so on.9–15 However, Yb3+ usually has no direct absorptions in UV-Vis (300–800 nm) range since it only has a single excited state 2F5/2 approximately 10000 cm−1 above the ground state 2F7/2. Therefore, Re3+ (Eu3+, Tb3+, Dy3+ and so on) is selected as a sensitizer in Re3+–Yb3+ co-doped phosphor. But this kind of materials usually have low absorption efficiency in UV-Vis light range, due to the forbidden 4f–4f transitions by the parity selection rule.
As we know, Pr3+ ion has the ability to absorb visible photons and emits NIR photons, because of its abundant energy levels in UV-Vis-NIR range. For instance, Pr3+ ion doped phosphors can be excited at the range of 440–700 nm due to the transitions from 3H4 to 1D2 (∼600 nm), 3PJ (J = 0, 1, 2) and 1I6 (440–490 nm), and emit NIR phonons at ∼1000 nm ascribed to the 3P0 → 1G4 and/or 1G4 → 3H4, matching well with the optimal spectral response of Si solar cells.16–19 To enhance the absorption of Pr3+ in ultraviolet region, here we explored a promising light conversion phosphor LiBaPO4:Eu2+, Pr3+ for Solar Spectral Convertor. It can exhibit an intense NIR emission of Pr3+ ion at about 1000 nm upon excitation with allowed 4f–5d absorption of Eu2+ ions.
The photoluminescence (PL) and photoluminescence excitation (PLE) spectra from 3 K to room temperature as well as the decay curves were measured by FSP920 Time Resolved and Steady State Fluorescence Spectrometers (Edinburgh Instruments) equipped with a 450 WW Xe lamp, a 100 W μF920H lamp, TM300 excitation monochromator and double TM300 emission monochromators, red sensitive PMT and R5509-72 NIR-PMT in a liquid nitrogen cooled housing (Hamamatsu Photonics K.K). The spectral resolution for the steady measurements is about 0.05 nm in UV-vis and about 0.075–0.01 nm in NIR, and the experimental conditions for the transient measurements are a pulse width of 1–2 μs, a repetition rate of 50 Hz and the lifetime range of 100 μs–200 s. For PL and PLE measurements at 3 K, the sample was mounted in a Optistat AC-V12 actively cooled optical cryostat, based on 0.25 W @ 4 K PTR (pulse tube refrigerator), with a ITC503 temperature controller and a water cooled compressor.
Fig. 1 XRD patterns of LiBaPO4:Eux2+, Pry3+ (x = 0, y = 0; x = 0.005, y = 0, 0.005; x = 0, y = 0.01) (a), and schematic diagram of the structure of LiBaPO4 along the a- (b) and c-direction (c). |
In order to clarify the assignment of the emissions peaks at 599 nm and 606 nm as shown in Fig. 2a, the luminescence decay curves at 3 K (λex = 483 nm and 599 nm, λem = 606 nm) of LiBaPO4:Pr0.013+ were measured and presented in Fig. 2d. Generally, the decay time of the emission from 1D2 is 50–250 μs, longer than that from 3P0, which is about 0.135–50 μs.22 It is clear that the decay curves have different characteristics and the average lifetimes are calculated to be about 84.8 μs and 3.8 μs when excited to 3P0 and 1D2 level, respectively. Therefore, it is reasonable to assign the emission peaks at 599 nm and 606 nm to 1D2 → 3H4 and 3P0 → 3H6 transitions.
However, why can the emission from 1D2 to 3H4 be detected when excited into 3P2 level? The energy transfer through multi-phonon relaxation and cross-relaxation process may give a clear interpretation. Fig. 3 presents the schematic diagrams of Pr3+ in LiBaPO4. It is widely accepted that the electrons excited to 3P2 level (2a) are able to relax to 3P0 level assisted by phonons (process 2b), because of the small energy difference between 3P2 and 3P0 level. There are two possible non-radiative path to generate the 1D2 → 3H4 transition (process 2c and 2d). One is the non-radiative relaxation from 3P0 to 1D2 level assisted by phonons, which is labeled as process 2c. The energy difference between 3P0 and 1D2 level is about 3697 cm−1. The maximum vibration frequency of phosphate is about 1037 cm−1.23 It means that three phonons are required to bridge this energy gap. Generally, phonons assisted non-radiative relaxation process from the upper level to the lower level is possible, if the energy gap is less than 5 times the energy of the highest energy phonon. Therefore, it is expected that multi-phonon relaxation process from 3P0 to 1D2 level may occur in LiBaPO4:Pr3+. The other is cross-relaxation process (2d), [3P0, 3H4] → [3H6, 1D2]. The energy gaps between 3P0 and 3H6 level, and 1D2 and 3H4 level are 16502 cm−1 and 17007 cm−1, respectively. The difference between them is only about 505 cm−1. Therefore, the cross-relaxation [3P0, 3H4] → [3H6, 1D2] (2d) is also efficient. In a word, the whole blue-to-red ET process could be expressed as (2a → 2b → 2c/2d → 599 nm) for the red emissions from the 1D2 level and (2a → 2b → 606 nm) from the 3P0 level.
To further study the effect of phonon to luminescence property of Pr3+ in LiBaPO4, Fig. 4a presents the normalized emission spectra of LiBaPO4:Pr0.013+ at different temperature. It's interesting to observe the temperature dependence of 3P1 → 3H5 transition peaked at 523 nm when excited into 3P0 level. At 3–100 K, Pr3+ ion gives no emission at 523 nm whereas above 100 K the 3P1 → 3H5 transition is clearly observed. Moreover, its emission intensity regularly increases with the increase of temperature. This phenomenon is due to the increase of electron–phonon interaction, which leads to more and more electrons populated from 3P0 up to 3P1 level with the rise of the temperature. In addition, the intensity ratio of 1D2 → 3H4 to 3P0 → 3H6 increases gradually with temperature increases, as shown in Fig. 3b, which strongly proves cross-relaxation or multi-phonon relaxation process (Fig. 3, 2d and c) from 3P0 to 1D2 level is efficient especially at room temperature. However, it is very hard to make a decision which one is dominant between multi-phonon relaxation process (2c) and cross-relaxation process (2d).
The energy difference between transition 1G4 → 1D2 (7203 cm−1) and 3H4 → 3F4 (6854 cm−1), is calculated to be only about 349 cm−1, as shown in Fig. 3. Thus, the cross-relaxation (2e), [1D2, 3H4] → [1G4, 3F4] will happen due to the small energy mismatch. In other words, electrons populate to the 1G4 level by two steps of cross-relaxation (2d–2e) when excited into 3P0 level. Consequently, it is theoretically expected that the NIR1 emissions from the 1G4 level could be observed. Fig. 5a presents the emission spectra of LiBaPO4:Pr0.013+ in NIR region at room temperature. Up to 3P0 (483 nm) and 1D2 (590 nm) level, the strong NIR emission at 1020 nm is observed and attributed to the transition of 1G4 → 3H4. The above results further prove that the cross-relaxation (2e), [1D2, 3H4] → [1G4, 3F4], is occurred in LiBaPO4:Pr0.013+. Beside the prominent NIR1 emission at 1020 nm, Pr3+ ion also gives the NIR2 emissions at 976 nm and 993 nm, attributed to the transitions from different sublevels of 1D2 to 3F4 level. To sum up, the process of NIR emissions from 1G4 and 1D2 levels could be expressed as (2b → 2c/2d → 2e → NIR1 and 2b → 2c/2d → NIR2), as shown in Fig. 3.
Fig. 5 Normalized PLE (a) and PL spectrum (b), decay curves (c) and absorption spectrum (d) of LiBaPO4:Pr0.013+ at room temperature. |
For the sake of verifying the assignment of emission at 976 nm, 993 nm and 1020 nm, Fig. 5b shows normalized decay curves and lifetime of LiBaPO4:Pr0.013+ at room temperature. If the emission transitions from the different stark levels of a excited level to a ground level, it is expect to have the approximated lifetime when up to the same excited levels. The lifetimes of emission at 1020 nm are 0.125 ms and 0.136 ms excited to 3P0 level (483 nm) and 1D2 level (588 nm), respectively. Similarly, the lifetimes of emissions at 993 nm and 976 nm are 0.085 ms and 0.068 ms excited to 1D2 level (588 nm). That is to say the emissions at 993 nm and 976 nm belong to the same transition, different to the 1020 nm. It is clearly that the lifetime of emission at 1020 nm is longer than that of the 993 nm and 976 nm because of the two steps of cross-relaxation processes. Therefore, the emissions at 976 nm (or 993 nm) and 1020 nm are attributed to the transitions of 1D2 → 3F4 and 1G4 → 3H4, respectively.
According to the report of A. Meijerink,24 the absorption of a photon to the 3PJ or 1I6 levels would be followed by the emission of two NIR photons, whereas absorption to 1D2 would result in emission of only one NIR photon. If the area ratio (RE) of the 3PJ and 1I6 band to 1D2 band in the excitation spectrum is twice as large as the ratio (RA) similarly obtained in the absorption spectra, the quantum cutting will be efficient. Fig. 5c and d show the excitation spectrum and the absorption spectrum of LiBaPO4:Pr0.013+. It can be found in Fig. 5c and d that the RE and RA are 4.0 and 4.4, respectively. It means that the blue-to-NIR quantum cutting process from the 3P0 level does not occur in LiBaPO4.
As mentioned above, the ideal light conversion material for Si solar cells should utilize the UV-vis region of the solar spectrum and possess an intense NIR emission, matching well with the maximum spectral response (λ ≈ 1000 nm) of Si semiconductor. As shown in Fig. 5a and c, LiBaPO4:Pr3+ meets the requirement of excitation and emission characteristics of light conversion material. However, it can only absorb the photons in blue and red region of the solar spectrum. Unfortunately, the intensity of the excitation peaks is very weak because the 4f–4f absorption transitions are forbidden by the parity selection rule. In order to widen the utilization of the solar spectrum towards UV region, we further investigate the luminescent properties of LiBaPO4:Eu2+,Pr3+ and explore the possibility of Eu2+ ion as donor absorbing UV photons, which greatly enhances the NIR emission of Pr3+ ion through efficient energy feeding by allowed 4f–5d absorption of Eu2+ ion with high oscillator strength.
Fig. 6a exhibits the PLE and PL spectra of LiBaPO4:Eu0.0052+, which contains 4f–5d allowed transition of Eu2+ ions at 250–425 nm and a blue emission band at 480 nm.21 It is worth noting that the Eu2+ single doped LiBaPO4 has no emission band at NIR range. Fig. 6b gives the PLE and PL spectra of LiBaPO4:Eu0.0052+, Pr0.0053+. Monitoring the emission at 1020 nm, the PLE spectrum contains a broad band and some sharp peaks. Comparing the PLE spectra of Eu2+ or Pr3+ singly doped in LiBaPO4 phosphors (Fig. 6a and 2c), the intense broad band centered at 338 nm should be attributed to 4f–5d allowed transition of Eu2+ ions. The other peaks at 444 nm, 468 nm, 483 nm and 590 nm are due to f–f transitions of Pr3+ ions. Similar as Fig. 5a, the emission band peaked at 1020 nm is attributed to transition of 1G4 → 3H4 of Pr3+. The appearance of 4f → 5d transitions of Eu2+ ions in the PLE spectrum and the existence of the 1G2 → 3H4 transitions of Pr3+ ion in the PL spectrum of LiBaPO4:Eu0.0052+,Pr0.0053+ indicate that the energy transfer from Eu2+ to Pr3+ occurs, as shown in Fig. 3. It demonstrates that LiBaPO4:Eu2+, Pr3+ can harvest UV-Vis photons of the incident solar spectral and significantly reduce the charges thermalization of Si solar cells by UV/vis-to-NIR spectral modification. At the same time, we compared our material with Ca2BO3Cl:Ce0.0023+, Tb0.013+, Yb0.013+, which was promising as useful light convert material for solar cell.25 It is obviously that the emission intensity of our phosphor is about 5 times as intense as Ca2BO3Cl:Ce0.0023+, Tb0.013+, Yb0.013+ in NIR area. Therefore, it can be a promising light conversion material to increase the efficiency of Si solar cells.
This journal is © The Royal Society of Chemistry 2017 |