Introduction

With the gradual depletion of fossil fuel energy sources and the increasing pollution of the environment, the development of new energy sources has become of utmost importance1,2,3,4,5,6,7,8,9,10,11,12. The most promising new energy source is solar energy. However, for current solar cells, the photoelectric transfer cost is high, and the efficiency is low. This results in a large difference between the significant potential of solar energy and its actual utilization rate5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20. Through quantum cutting, a high-energy photon can be converted into many low-energy photons. It is a new method to reduce the losses in solar cells by modifying the distribution of the incident solar light energy, which can be used to generate solar energy more effectively5, 12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33. It is possible to apply the quantum cutting method to all types of solar cells without changing their structures. The ability of photovoltaic cells to convert sunlight into electricity makes them prime candidates for the effective large-scale capture and conversion of solar energy.

Green and Trupke originally proposed the theory of a “two-photon quantum cutting silicon solar cell” in 200210. They reported a maximum theoretical efficiency of 38% for such a device, and it exhibited sensitivity to solar light at wavelengths from 280 nm to 1100 nm10. Meijerink and Vergeer first demonstrated an experiment on the near-infrared, two-photon quantum cutting phenomenon in YbxY1−xPO4:Tb3+ phosphors in 20051, which was conducted after they reported a well-known visible quantum cutting experiment for an Eu3+/Gd3+ system in Science 2. Since 2007, several groups, including Meijerink1, 3, 19, 21, 27, Qiu and Zhou5, 12, 15, Wang and Chen6, 11, 13, Huang24, Zou and Wang25, Fedorov and Luginina9, Zhang7, 27, Xia and Hu29, Guo and Chen23, and Song and Tao30, 32, and more17, 20, 22, have published more than 200 articles on the second-order, near-infrared, quantum cutting luminescence phenomena of sensitizer-Yb3+ co-doped materials1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25, which were used to develop two-photon quantum cutting silicon solar cells. Near-infrared quantum cutting has become a hot topic in the field of science and nature. Recently, it was proven by several groups, including Hu and Hao22, Li17, and Song33, etc. that the actual photoelectric conversion efficiency of silicon based solar cells can be enhanced by the quantum cutting effect. Meijerink19, 27, Qiu and Zhou12, 15, Zhang7, 27, Huang24, and our group8, 28 have reported experimental research on first-order, multiphoton, near-infrared quantum cutting in Er3+ or Tm3+ activator-ion doped materials. This improvement has led to the development of first-order, multiphoton quantum cutting germanium (Ge) and silicon–germanium (Si–Ge) solar cells1, 7, 8, 12, 15, 27, 28, which are sensitive to wavelengths of 280–1850 nm and are environmentally friendly. Their maximum efficiency can clearly exceed 38%. On the other hand, near-infrared quantum cutting can be used in 1.8–2.0 μm near-to-mid infrared lasers. These types of lasers have potential applications in micro-surgery, tissue welding, range finding, remote sensing, environmental trace-gas detection, and biophysical applications12, 31.

One key factor has contributed to the important developments and improvements of first-order near-infrared quantum cutting, compared to second-order methods. Because there are no energy resonances between donors and acceptors and its energy of a donor equals two times the energy of an acceptor for second-order near-infrared quantum cutting, thus it only has a second-order process but not a first-order process. However, a first-order process is approximately 1000 times larger than a second-order process, as indicated by Meijerink3. In particular, for the first-order near-infrared multiphoton quantum cutting of Er3+ or Tm3+ ions, their cross energy transfers are excellent first-order processes with high oscillator intensities and small energy mismatches. Therefore, the first-order, near-infrared, multiphoton quantum cutting processes with Er3+ or Tm3+ activators have large rates, high efficiencies, and excellent prospects for different applications.

Results

Absorption

The absorption spectra of samples (A) and (B) are shown in Fig. 1. We found that the absorption peaks of sample (B) Tm3+(0.5%):telluride glass are positioned at 1699 nm, 1212 nm, 793 nm, 687 nm, and 465 nm. These absorption peaks are for the 3H6 → 3F4, 3H6 → 3H5, 3H6 → 3H4, 3H6 → 3F3, and 3H6 → 1G4 absorption transitions of the Tm3+ ions, respectively16, 18. We also found that the absorption peaks of the Er3+ ion of sample (A) Er(8.0%)Tm3+(0.5%):telluride glass are positioned at (1531 nm, 1497 nm), 977 nm, 795 nm, 653 nm, 544 nm, 522 nm, 489 nm, 451 nm, 408 nm, 380 nm, and 366 nm. These absorption peaks are for the 4I15/2 → 4I13/2, 4I15/2 → 4I11/2, 4I15/2 → 4I9/2, 4I15/2 → 4F9/2, 4I15/2 → 4S3/2, 4I15/2 → 2H11/2, 4I15/2 → 4F7/2, 4I15/2 → 4F5/2, 4I15/2 → 2H9/2, 4I15/2 → 4G11/2, and 4I15/2 → 4G9/2 absorption transitions of the Er3+ ions, respectively16, 18. The schematic diagram of the energy level structures for the Er3+ and Tm3+ ions are shown in Fig. 2.

Figure 1
figure 1

Absorption spectra of samples (A) Er3+(8%)Tm3+(0.5%):telluride glass and (B) Tm3+(0.5%):telluride glass.

Full size image
Figure 2
figure 2

Schematic diagrams of the energy-level structures of Er3+ and Tm3+ ions and the quantum cutting process.

Full size image

Excitation spectra

First, we selected the 1800-nm infrared luminescence wavelength of the Tm3+ ions in telluride glass to measure the infrared excitation spectra, from 250 nm to 850 nm, in sample (A) Er3+(8%)Tm3+(0.5%):telluride glass and sample (B) Tm3+(0.5%):telluride glass. The results are shown in Fig. 3. It was found that there are four excitation peaks for sample (B) Tm3+(0.5%):telluride glass, which are positioned at 358 nm, 467 nm, 686 nm, and 790 nm. Their excitation peak intensities are approximately 4.03 × 102, 5.59 × 102, 1.23 × 103, and 1.56 ×  × 103, respectively. These four excitation peaks are for the 3H6 → 1D2, 3H6 → 1G4, 3H6 → 3F3, and 3H6 → 3H4 transitions of the Tm3+ ion16, 18. It was also found that there are ten excitation peaks for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, which are positioned at 366 nm, 380 nm, 408 nm, 451 nm, 489 nm, 523 nm, 544 nm, 652 nm, 686 nm, and 795 nm. Their peak intensities are approximately 5.33 × 103, 6.63 × 103, 3.48 × 103, 3.03 × 103, 4.48 × 103, 5.16 × 103, 3.37 × 103, 3.29 × 103, 1.40 × 103, and 3.16 × 103, respectively. The excitation peaks at 366 nm, 380 nm, 408 nm, 451 nm, 489 nm, 523 nm, 544 nm, 652 nm, and 795 nm are for the 4I15/2 → 4G9/2, 4I15/2 → 4G11/2, 4I15/2 → 2H9/2, 4I15/2 → 4F5/2, 4I15/2 → 4F7/2, 4I15/2 → 2H11/2, 4I15/2 → 4S3/2, 4I15/2 → 4F9/2, and 4I15/2 → 4I9/2 transitions of the Er3+ ion, respectively16, 18. The excitation peak at 686 nm is for the 3H6 → 3F3 transition of the Tm3+ ions. The peak excitation intensity of 6.63 × 103 for the 380 nm 4I15/2 → 4G11/2 transition of the Er3+ ions is, respectively, 16.5, 11.9, 5.4, and 4.3 times larger than the peak excitation intensities of 4.03 × 102, 5.59 × 102, 1.23 × 103, and 1.56 × 103 for the 3H6 → 1D2, 3H6 → 1G4, 3H6 → 3F3, and 3H6 → 3H4 transitions of the Tm3+ ions.

Figure 3
figure 3

Infrared excitation spectra of samples (A) Er3+(8%)Tm3+(0.5%):telluride glass and (B) Tm3+(0.5%):telluride glass when monitored at 1800 nm for the 3F4 → 3H6 luminescence of the Tm3+ ions.

Full size image

Then, we selected the 651-nm visible luminescence wavelength of Tm3+ ions in telluride glass to measure the visible excitation spectra, from 250 nm to 600 nm, in sample (A) Er3+(8%)Tm3+(0.5%):telluride glass and sample (B) Tm3+(0.5%):telluride glass. The results are shown in Fig. 4(a). From Fig. 4(a), the excitation spectrum of sample (B) Tm3+(0.5%):telluride glass, when monitored at the 651 nm visible luminescence wavelength, has two excitation peaks, which are positioned at 359.0 nm and 465.0 nm with peak intensities of 1.16 × 105 and 6.02 × 105, respectively. The 359.0 nm and 465.0 nm excitation peaks are attributed to the 3H6 → 1D2 and 3H6 → 1G4 transitions of the Tm3+ ions16, 18, respectively. However, the excitation spectrum of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, when monitored at the 651 nm visible luminescence wavelength, has approximately seven excitation peaks, which are positioned at 365.5 nm, 378.0 nm, 406.5 nm, 450.5 nm, 488.5 nm, 520.5 nm, and 544.5 nm with excitation peak intensities of 1.02 × 105, 1.52 × 105, 4.76 × 104, 3.75 × 104, 6.67 × 104, 1.01 × 105, and 3.15 × 104, respectively. These excitation peaks correspond to the 4I15/2 → 4G9/2, 4I15/2 → 4G11/2, 4I15/2 → 2H9/2, 4I15/2 → 4F5/2, 4I15/2 → 4F7/2, 4I15/2 → 2H11/2, and 4I15/2 → 4S3/2 transitions of the Er3+ ions, respectively16, 18. This implies that the excitation energy for the 651 nm luminescence in sample (A) Er3+(8%)Tm3+(0.5%):telluride glass came from the Er3+ ions already. We also selected the 556 nm visible luminescence wavelength of Er3+ ions in the telluride glass to measure the visible excitation spectra, from 250 nm to 535 nm, in sample (A) Er3+(8%)Tm3+(0.5%):telluride glass and sample (C) Er3+(0.5%):telluride glass. These results are shown in Fig. 4(b). The spectrum of Fig. 4(b) is the characteristic excitation spectrum of Er3+ ions16, 18.

Figure 4
figure 4

(a) The visible excitation spectra of samples (A) Er3+(8%)Tm3+(0.5%):telluride glass and (B) Tm3+(0.5%):telluride glass when monitored at 651 nm for the 4F9/2 → 4I15/2 luminescence of Er3+ ions and 1G4 → 3F4 luminescence of Tm3+ ions. (b) The visible excitation spectra of samples (A) Er3+(8%)Tm3+(0.5%):telluride glass and (C) Er3+(0.5%):telluride glass when monitored at 556 nm for the 4S3/2 → 4I15/2 luminescence of the Er3+ ions.

Full size image

Luminescence spectra

First, we selected the 4I15/2 → 4G11/2 absorption wavelength, 380 nm, of the Er3+ ions in sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, and the 3H6 → 1D2, 3H6 → 1G4, 3H6 → 3F3, and 3H6 → 3H4 absorption wavelengths of 358 nm, 467 nm, 686 nm, and 790 nm of the Tm3+ ions in sample (B) Tm3+(0.5%):telluride glass as the excitation wavelengths to measure the infrared luminescence spectra at wavelengths from 1200 nm to 2800 nm. The results are shown in Fig. 5(a). There are two luminescence peaks for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, which are positioned at 1537 nm and 1800 nm. These luminescence peaks are the 1537 nm 4I13/2 → 4I15/2 luminescence of the Er3+ ions, and the 1800 nm 3F4 → 3H6 luminescence of the Tm3+ ions16, 18. The luminescence peak intensities of the 1537 nm and 1800 nm peaks are approximately 5.32 × 102 and 1.73 × 103, respectively. There are two luminescence peaks for sample (B) Tm3+(0.5%):telluride glass, which are positioned at 1468 nm and 1800 nm. These two luminescence peaks are for the 1468 nm 3H4 → 3F4 luminescence of the Tm3+ ions and the 1800 nm 3F4 → 3H6 luminescence of the Tm3+ ions16, 18. The luminescence peak intensities of the 1800 nm peaks, when the sample is excited by 358 nm, 467 nm, 686 nm, and 790 nm light, are approximately 8.90 × 101, 1.45 × 102, 3.10 × 102, and 3.96 × 102, respectively. The 1800-nm luminescence peak intensity, 1.73 × 103, of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass when excited by 380 nm light is approximately 19.5, 12.0, 5.6, and 4.4 times larger than that of sample (B) Tm3+(0.5%):telluride glass when excited by 358 nm, 467 nm, 686 nm, and 790 nm light, respectively.

Figure 5
figure 5

(a) Infrared luminescence spectra for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass when excited by 380 nm, 358 nm, 467 nm, 686 nm, and 790 nm light for the 4I15/2 → 4G11/2 absorption of the Er3+ ions, 3H6 → 1D2, 3H6 → 1G4, 3H6 → 3F3, and 3H6 → 3H4 absorption of Tm3+ ions. (b) Infrared luminescence spectra for sample (A)Er3+(8%)Tm3+(0.5%):telluride glass when excited by 380 nm, 408 nm, 522 nm, 544 nm, 652 nm, and 795 nm light for the 4I15/2 → 4G11/2, 4I15/2 → 2H9/2, 4I15/2 → 2H11/2, 4I15/2 → 4S3/2, 4I15/2 → 4F9/2, 4I15/2 → 4I9/2 absorption of the Er3+ ions.

Full size image

We then selected the 4I15/2 → 4G11/2, 4I15/2 → 2H9/2, 4I15/2 → 2H11/2, 4I15/2 → 4S3/2, 4I15/2 → 4F9/2, and 4I15/2 → 4I9/2 absorption wavelengths of 380 nm, 408 nm, 522 nm, 544 nm, 652 nm, and 795 nm for the Er3+ ions in sample (A) Er3+(8%)Tm3+(0.5%):telluride glass as the excitation wavelengths to measure the infrared luminescence spectra, from 1200 nm to 2800 nm. The results are shown in Fig. 5(b). Their luminescence peak intensities are approximately 1.73 × 103, 6.53 × 102, 1.38 × 103, 7.83 × 102, 8.48 × 102, and 8.17 × 102, respectively.

In addition, we selected the 4I15/2 → 4G11/2 absorption wavelength, 380 nm, of the Er3+ ions as the excitation wavelength to measure the infrared luminescence spectra, from 1200 nm to 2800 nm, for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass and sample (C) Er3+(0.5%):telluride glass. The results are shown in Fig. 6. There is only one main luminescence peak for sample (C) Er3+(0.5%):telluride glass, which is positioned at 1537 nm. This luminescence peak is the 1537 nm 4I13/2 → 4I15/2 transition of the Er3+ ions16, 18. Its luminescence peak intensity is approximately 9.78 × 102. The ratio of the 1800-nm luminescence peak intensity of 1.73 × 103 of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, to the 1537-nm luminescence peak intensity of 9.78 × 102 of sample (C) Er3+(0.5%):telluride glass, is approximately 1.8. Meanwhile, the ratio of the 1800-nm luminescence integral area intensity of 4.76 × 105 for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, to the 1537-nm luminescence integral area intensity of 9.55 × 104 for sample (C) Er3+(0.5%):telluride glass, is approximately 5.0. From the results of Figs 5(a) and 6, we can conclude that the infrared luminescence intensity of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, is much larger than that of sample (B) Tm3+(0.5%):telluride glass or sample (C) Er3+(0.5%):telluride glass.

Figure 6
figure 6

Visible and infrared luminescence spectra of samples (A) Er3+(8%)Tm3+(0.5%):telluride glass and (C) Er3+(0.5%):telluride glass when excited by 380 nm light for the 4I15/2 → 4G11/2 absorption of Er3+ ions.

Full size image

Finally, we selected the 4I15/2 → 4G11/2 absorption wavelength, 380 nm, of the Er3+ ions as the excitation wavelength to measure the visible luminescence spectra, from 395 nm to 728 nm, for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass and sample (C) Er3+(0.5%):telluride glass. The results are also shown in Fig. 6. There are four luminescence peaks, which are positioned at 408.0 nm, 525.0 nm, (545.0 nm/556.0 nm), and 658.0 nm. These four luminescence peaks are for the 2H9/2 → 4I15/2, 2H11/2 → 4I15/2, 4S3/2 → 4I15/2, and 4F9/2 → 4I15/2 luminescence transitions of the Er3+ ions, respectively. Their luminescence peak intensities are approximately 3.56 × 104, 2.25 × 104, 8.62 × 104, and 6.05 × 103, respectively, for sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, and are approximately 4.37 × 104, 1.10 × 106, 2.39 × 106, and 2.58 × 104, respectively, for sample (C) Er3+(0.5%):telluride glass.

Lifetime dynamics

We used a 380-nm pulsed light from a xenon lamp as the excitation light source to measure the lifetimes of the 1537-nm luminescence peak in samples (A) Er3+(8%)Tm3+(0.5%):telluride glass, (C) Er3+(0.5%):telluride glass, and (D) Er3+(8.0%):telluride glass. The measured results are shown in Fig. 7.

Figure 7
figure 7

Lifetimes of the 1537-nm luminescence peaks in samples (A) Er3+(8%)Tm3+(0.5%):telluride glass, (C) Er3+(0.5%):telluride glass, and (D) Er3+(8.0%):telluride glass when excited by 380 nm light for the 4I15/2 → 4G11/2 absorption of Er3+ ions.

Full size image

According to the literature in the field of infrared quantum cutting, the efficiency of energy transfer among Er3+ ions can be calculated by using formula (1) below. Similarly, the efficiency of energy transfer between Er3+ and Tm3+ ions can be calculated by using formula (2)1, 3, 5,6,7,8,9,10,11,12, 15, 20,21,22,23,24,25,26,27,28:

$${\eta }_{tr,x \% Er}\approx 1-\frac{\int {I}_{x \% Er}dt}{\int {I}_{{\rm{0.5}} \% Er}dt}$$
(1)
$${\eta }_{tr,x \% Er,y \% Tm}\approx 1-\frac{\int {I}_{x \% Er,y \% Tm}dt}{\int {I}_{x \% Er}dt}$$
(2)

where I denotes the light intensity, x%Er represents the concentration of Er3+ ions, and y%Tm represents the concentration of Tm3+ ions. It is assumed that the energy transfer between Er3+ ions is negligible when x = 0.5%. Therefore, \({I}_{0.5 \% Er}\) can represent the case of non-energy transfer. It is known that there is an intense resonant energy diffusion {4I13/2(Er3+) → 4I15/2(Er3+), 4I15/2(Er3+) → 4I13/2(Er3+)} among Er3+ ions in (A) Er3+(8%)Tm3+(0.5%):telluride glass and (D) Er3+(8.0%):telluride glass, because both the concentrations of Er3+ ions and the populations in the 4I13/2(Er3+) first-excited state are very high. Therefore, resonant energy transfer to nearby Er3+ ions will be large. The excitations will lose their energy to impurity, defects, or trap states – a process referred to as concentration quenching. Formula (1) calculates only the efficiency of the resonant energy transfer.

From Fig. 7, we can calculate the integrated sum value for the luminescence lifetime curves of samples (A) Er3+(8%)Tm3+(0.5%):telluride glass, (C) Er3+(0.5%):telluride glass, and (D) Er3+(8.0%):telluride glass. The results are:\(\int I{(1537nm)}_{0.5 \% Er}dt=4.886\), \(\int I{(1537nm)}_{8.0 \% Er}dt=2.892\), and \(\int I{(1537nm)}_{8 \% Er0.5 \% Tm}dt=0.874\). According to formula (1), we obtained the efficiency of resonant energy transfer {4I13/2 → 4I15/2, 4I15/2 → 4I13/2} among the Er3+ ions as follows: \({\eta }_{tr,8.0 \% Er}(1537nm)=40.8 \% \). From formula (2), we obtain the efficiency of energy transfer {4I13/2(Er3+) → 4I15/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)} between the Er3+ and Tm3+ ions as follows: \({\eta }_{tr,8 \% Er,0.5 \% Tm}=1-\frac{\int {I}_{8 \% Er,0.5 \% Tm}dt}{\int {I}_{8 \% Er}dt}=69.8 \% \).

Figure 8 shows the infrared luminescence of Er3+ ions and the infrared absorption of Tm3+ ions. We found that there is an obvious overlap between the infrared luminescence of Er3+ donor ions and the infrared absorption of Tm3+ acceptor ions. This results in a very strong energy transfer {4I13/2(Er3+) → 4I15/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)} between the Er3+ and Tm3+ ions. The transfer efficiency of \({\eta }_{tr,8 \% Er,0.5 \% Tm}=69.8 \% \) is very reasonable for photovoltaic applications.

Figure 8
figure 8

The infrared luminescence of sample (C) Er3+(0.5%):telluride glass when excited by 380-nm light for the 4I15/2 → 4G11/2 absorption of Er3+ ions and the infrared absorption of sample (B) Tm3+(0.5%):telluride glass.

Full size image

Discussion

From the results of the measurements shown in Figs 5 and 6 and their analyses, we found that the near-infrared 1800-nm luminescence intensity of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass is approximately 4.4 to 19.5 times larger than that of sample (B) Tm3+(0.5%):telluride glass, and is approximately 5.0 times larger than that of sample (C) Er3+(0.5%):telluride glass. Meanwhile, from Fig. 6, the visible luminescence intensity of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass is much smaller than that of sample (C) Er3+(0.5%):telluride glass. Moreover, from Figs 3 and 4, we found that the excitation spectra of the 1800-nm infrared luminescence and the 522 nm and 652 nm visible luminescences of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass are very similar to those of the Er3+ ions of sample (C) Er3+(0.5%):telluride glass, with respect to the shape of the spectral waveforms and peak wavelengths. From Figs 7 and 8, we found an overlap and energy transfer between the infrared luminescence of the Er3+ donor ions and the infrared absorption of the Tm3+ acceptor ions. The efficiency of the first-order energy transfer {4I13/2(Er3+) → 4I15/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)} between the Er3+ and Tm3+ ions is approximately \({\eta }_{tr,8 \% Er,0.5 \% Tm}=69.8 \% \). From our previous work28, we know that a telluride glass with an 8% mol concentration of Er3+ ions will exhibit intense first-order near-infrared quantum cutting luminescence phenomena. It is obvious that there is no cross-energy transfer for samples (B) Tm3+(0.5%):telluride glass and (C) Er3+(0.5%):telluride glass, because their 0.5% concentration of rare earth ions is low. However, sample (A) Er3+(8%)Tm3+(0.5%):telluride glass has strong cross-energy transfer between Er3+ ions, because its 8% concentration of Er3+ ions is high. Therefore, we can conclude that the observed behaviour is an important multiphoton first-order near-infrared quantum cutting luminescence phenomenon of novel Er3+/Tm3+ ion pairs. Therefore, sample (A) Er3+(8%)Tm3+(0.5%):telluride glass first exhibits an intense first-order near-infrared quantum cutting among Er3+ ions, and then, the energy is transferred from Er3+ ions to Tm3+ ions. This results in the intense multiphoton first-order near-infrared quantum cutting 1800-nm luminescence of the Tm3+ ions.

The schematic diagrams of the energy-level structures of Er3+ and Tm3+ ions and the quantum cutting process are shown in Fig. 2.

When the 4G11/2 energy level is excited by 380-nm light, many Er3+ ions may populate at the 4G11/2 energy level because its absorption is very strong. The Er3+ ions undergo an intense {4G11/2  →  4I13/2, 4I15/2 → 2H11/2} ETr101-ETa06 first-order cross-energy transfer process. The transition mismatch, ΔE = 623 cm−1, is moderate, but the reduced matrix elements (Uλ)2 (0.1005, 0.2648, 0.2570) and (0.7158, 0.4138, 0.0927) of the Er3+ ions are very large16, 18, and the multiphonon non-radiative relaxation is moderate, therefore the first-order cross-energy transfer rate of {4G11/2 → 4I13/2, 4I15/2 → 2H11/2} ETr101-ETa06 is large. The population of the 4G11/2 energy level may be initially transferred to the first excited state 4I13/2 and the 2H11/2 energy levels mainly through the {4G11/2 → 4I13/2, 4I15/2 → 2H11/2} ETr101-ETa06 first-order cross-energy transfer process. The population in the 2H11/2 energy level may be sequentially transferred to 4I13/2 via {2H11/2 → 4I9/2, 4I15/2 → 4I13/2} ETr63-ETa01 and {4I9/2 → 4I13/2, 4I15/2 → 4I13/2} ETr31-ETa01. This would result in the intense four-photon first-order near-infrared quantum cutting of the 4I13/2 → 4I15/2 luminescence. Moreover, the 3F4 level of Tm3+ ions is positioned at a slightly lower energy than the 4I13/2 level of Er3+ ions. There is a significant overlap and a first-order energy transfer {4I13/2(Er3+) → 4I15/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)} between the Er3+ and Tm3+ ions. Furthermore, the back-energy transfer {3F4(Tm3+) → 3H6(Tm3+), 4I15/2(Er3+) → 4I13/2(Er3+)} might be relatively very small, since it is an anti-Stokes process. Therefore, this would result in the very intense multi-photon first-order quantum cutting 1800-nm luminescence of the Tm3+ ions. This is the main cross-energy transfer process, which is shown in the Fig. 2.

Meanwhile, there are the subordinate cross-energy transfer processes occurred between Er3+ and Tm3+ ions directly. For example, {4S3/2(Er3+) → 4I9/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)} first-order cross-energy transfer process is large also even the concentration of Er3+ and Tm3+ is 8% and 0.5%. Because its transition mismatch, ΔE = −82 cm−1, is small, its reduced matrix elements (Uλ)2 (0, 0.0765, 0.2569) and (0.5375, 0.7261, 0.2382) of the Er3+ and Tm3+ ions are large16, 18, and its multiphonon non-radiative relaxation is small. The population of the 4S3/2 energy level may be initially transferred to the excited state 4I9/2(Er3+) and the 3F4(Tm3+) energy levels mainly through the {4S3/2(Er3+) → 4I9/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)} first-order cross-energy transfer process. The population in the 4I9/2(Er3+) energy level may be sequentially transferred to 3F4(Tm3+) via {4I9/2 → 4I13/2, 4I15/2 → 4I13/2} ETr31-ETa01 and {4I13/2(Er3+) → 4I15/2(Er3+), 3H6(Tm3+) → 3F4(Tm3+)}. This would result in the intense three-photon first-order near-infrared quantum cutting 1800-nm luminescence of the Tm3+ ions.

As we know, the energy band gap of GaN materials is approximately 3.4 eV, which corresponds to 27423 cm−1 (365 nm) light. It is easy to achieve very intense 380-nm luminescence in a GaN light emitting diode (LED)26. Therefore, using the excellent quantum cutting phenomenon in the novel Er3+/Tm3+ ion pair to construct a near-to-mid infrared (1.8–2.0 μm) laser pumped by a GaN LED is a significant and useful prospect12, 31. Quantum cutting, GaN LEDs, and near-to-mid infrared (1.8–2.0 μm) lasers are all currently hot topics in field of science and nature. It is possible to significantly enhance the properties of these lasers12, 31. To the best of our knowledge, the present manuscript is the first to report the first-order quantum cutting effect of an Er3+-Tm3+ ion pair.

To summarize, we measured the absorption, excitation, and luminescence spectra, as well as the lifetime dynamics of Er3+/Tm3+ co-doped telluride glasses. An interesting multiphoton near-infrared quantum cutting luminescence phenomenon from novel Er3+-Tm3+ ion pairs was found. This can facilitate the development of next-generation environmentally friendly germanium solar cells. In addition, using the excellent quantum cutting ability of our novel Er3+/Tm3+ ion pairs to construct a near-to-mid infrared (1.8–2.0 μm) laser pumped by GaN LEDs is a promising prospect.

Methods

Synthesis

The samples used in the present work were (A) Er3+(8%)Tm3+(0.5%):telluride glass, (B) Tm3+(0.5%):telluride glass, (C) Er3+(0.5%):telluride glass, and (D) Er3+(8.0%):telluride glass. The composition of sample (A) Er3+(8%)Tm3+(0.5%):telluride glass, for example, was 70TeO2–25ZnO-5La2O3–8Er2O3–0.5Tm2O3. The telluride glasses were manufactured using highly purified TeO2, ZnO, La2O3, Er2O3, and Tm2O3 powders as the starting materials. The well-mixed raw materials were placed in an alumina crucible. The samples were melted at 900 °C for 50 min under an oxygen atmosphere. A dry oxygen atmosphere was introduced to remove hydroxyl groups. The melts were then poured into a preheated stainless steel mould and annealed for several hours near the glass transition temperature, Tg (approximately 300 °C). The annealed samples were cut and polished to a size of 16 mm × 20 mm × 5.5 mm for optical measurements.

Characterization

The equipment used in our experiment was a FL3-2iHR fluorescence spectrometer (Horiba-JY Co., America, Japan, and France). The excitation light source was a xenon lamp. The visible light detector was an R2658p photomultiplier. The infrared detector was a DSS-PS020T PbS detector. For all results, the signal intensities at the same wavelength in the same figure can be compared directly. The absorptions were measured using a UV3600 spectrophotometer (Shimadzu, Japan). The lifetime dynamics were recorded using the same fluorescence spectrometer, with an excitation wavelength of 378 nm, a measurement range of 22 ms, a peak present of 2.80 × 104, a sweep present of 1.28 × 105, and a delay of 0%.