Introduction

Recently, Er3+ ions have been widely investigated and used in the field of optics as active optical ions because of their unique luminescence properties and low-cost, high-efficiency and commercially available the pump resources1,2,3,4,5,6,7,8. In particular, Er3+ ions can easily achieve population inversion and amplification around 2.7 μm because of the narrow energy gap (~3600 cm−1) between 4I11/2 and 4I13/29. This radiation is interesting for many medical and spectroscopic applications10, optical communication2 and as an excitation source for infrared (IR) optical parametric oscillators9. However, the efficiency of the 2.7 μm emissions is strongly dependent on the phonon vibration of the medium because the mid-infrared (2.6–2.9 μm) fluorescence will be quenched in a high phonon-energy environment11. Therefore, an appropriate host material is extremely crucial for achieving high efficiency mid-infrared emissions.

Mid-infrared host materials are mainly low-phonon-energy fluoride materials, such as fluoride glasses and fluoride single crystals. Among these host materials, Er3+-doped fluoride glasses have achieved efficient 2.7 μm lasing emissions, but exhibit inferior thermal and mechanical properties6,12. Er3+-doped fluoride single crystals are considered ideal host materials for Er3+ ions while retaining their excellent 2.7 μm emission performance because of their low phonon energy, high thermal conductivity and excellent heat shock resistance9. Unfortunately, the strict requirements for fabricating large, homogenously doped and high-quality fluoride single crystals is a serious obstacle that hinders the crystal-based mid-infrared applications9,13.

Transparent glass ceramics (GCs), which possess the combined properties of glasses and crystals, are promising optical materials that compete with single crystals and glasses. Meanwhile, many experimental studies have confirmed that rare-earth (RE) ions are selectively embedded in the fluoride crystals distributed in the glass matrix. Therefore, Er3+-doped GCs containing fluoride crystals are expected to achieve highly efficient mid-infrared emissions. Wei6 and Wu12 have successfully observed efficient 2.7 μm emissions in Er3+-doped NaYF4 and CaF2 nanocrystals embedded oxyfluoride GCs prepared by melt-quenching method, respectively. As is well known, some disadvantages of melt-quenching method, such as high preparation temperature and energy consumption, uncontrollable crystal formation and long and complex heat treatment, severely hinder the GCs-based applications14,15,16. On the other hand, the optical properties in these GCs are achieved only through a careful partial crystallization of the nanometer scale grains. However, network modifiers and impurities in the glass matrix tend to accumulate on the surface of the nanocrystals because of the high surface energy, which may change the properties of the nanocrystals17. Meanwhile, defect energy levels are easily introduced into the forbidden band of nanocrystals because of large surface defects, resulting in defect luminescence18. Large grains can prevent the problems encountered in nanocrystals because of a small surface-to-volume ratio. Fan et al.9 obtained intense IR fluorescence around 2.7 μm in Er3+-doped transparent GC containing large CaF2 crystals by a technique similar to the liquid-phase-sintering method, which implies that Er3+-doped transparent GCs containing large fluoride grains exhibit a high potential for mid-infrared applications. Additionally, abovementioned method possesses low preparation temperature and precise crystal phase control that compete with melt-quenching method. Unfortunately, a complex and time-consuming preparation process including coprecipitation, ball-milling and highly precise temperature and time control is requirement. Most importantly, long-time sintering is necessary for abovementioned method to remove pores and achieve the growth of crystals, which causes strong glass components volatilization. The changes of glass components will increase the differences in the refractive index between crystalline phase and glass phase and scattering losses. Meanwhile, glass components volatilization entails a certain degree of randomness, which is difficult to accurately adjust the refractive index of glass matrix. Measures must be taken to address these problems.

In order to avoid the difficulties of6,9,12, a liquid-phase sintering route was designed by tuning components of PG and optimizing the preparation process in foundation of existing technique9,19, which is directly sintering a mixture of commercial fluoride powders and PG powders prepared by grinding at low temperature for a short time. This liquid-phase-sintering route has several advantages over existing GCs preparation method such as: precise control over the crystal phase and refractive index, short preparation time, low preparation temperature and energy consumption and simple and low-cost synthetic approaches, which results in a high financial and energetic saving.

In this study, based on remarkable optical performances and high potential for mid-infrared applications of SrF2 crystal and GCs containing SrF2 crystals20,21,22, Er3+-doped transparent GCs containing micron-sized SrF2 crystals were fabricated by abovementioned liquid-phase sintering route and characterized. To the best of our knowledge, we obtained, for the first time, transparent GCs containing micron-sized SrF2 crystals and 2.7 μm emissions in GCs containing SrF2. Intense 2.7 μm emissions generated by GCs have been discussed in detail for possible applications in mid-infrared lasers and amplifiers.

Experimental

Sample preparation

Fluorophosphate (FP) glass, which is a promising laser material, has remarkable spectral properties, good chemical and mechanical stability, low OH– content and a low melting temperature23. In particular, the refractive index of FP glass can be readily tuned to match that of SrF2, yielding highly transparent GCs containing large SrF2 crystals. In our experiment, Er3+ doped FP glass with a nominal molar composition of 35CaF2-18MgF2-10SrF2-18BaF2 -3NaF-6AlF3- 7Na(PO3)3- 2.7Al(PO3)3 -0.3ErF3, denoted as precursor glass (PG), was prepared by the conventional melting-quenching technique. The refractive index of FP glass was matched with that of the SrF2 crystal using the GE-SYSTEM software. Samples were prepared using analytical-grade NaF, MF2 (M:Ca, Sr, Ba), AlF3, ErF3, Na(PO3)2 and Al(PO3)3 by the liquid-phase sintering method. A GC with a composition of 45SrF2-55 PG (wt%) was prepared from SrF2 and PG powders. SrF2 powders (4.5 g) were blended directly with PG powders (5.5 g) by grinding in an agate mortar for 5 min. The mixture was melted in an electric furnace at 820 °C for 15 min. The melts were then quenched in stainless steel molds and pressed into thin plates. The samples were annealed at 380 °C (Tg = 392 °C) for 2 h to reduce internal stress, when Er3+:SrF2 GCs were obtained.

Sample characterization

The crystallization behavior of the samples was investigated using a Philips PANalytical X’Pert X-ray diffraction (XRD) system at 40 kV and 40 mA with a Cu target X-ray tube. The transmittance characteristics were investigated by a Lambda 1050 UV/VIS/NIR spectrophotometer (Perkin Elmer) for wavelengths ranging from 300 to 3000 nm. The thermal behavior of the PG was characterized using thermogravimetry (TG) and differential thermal analysis (DTA) at a heating rate of 5 K/min in air from 100 °C to 1000 °C using an EXSTAR SII:TG/DTA7300 from Japan. The refractive index of the samples was obtained by the prism minimum deviation method using a WYV V-Prism refractometer (CANY, Shanghai, China) with an accuracy of nD ± 5 × 10−5. The microstructural evolution of the samples was studied by a field-emission scanning electron microscope (FESEM, HitachiS-4800, Japan) equipped with an energy-dispersive X-ray spectrometer (EDS) and transmission microscope (Olympus). The emission spectrum with excitation of a diode laser at 980 nm was recorded using by a PMT detector (R928) at 25 °C.

Results and Discussion

XRD measurements were performed to examine the crystallization behavior. Fig. 1 shows the XRD pattern of the Er3+-doped PG, Er3+-doped GC and SrF2 powders, respectively. Only two broad humps are observed in the XRD pattern of the PG, indicating that no spontaneous crystallization occurs in PG. After liquid-phase sintering at 820 °C for 15 min, two broad envelopes of glass diffraction and some intense sharp peaks of the crystalline SrF2 cubic phase emerge in the XRD curve of GC, which confirms that SrF2 crystals were successfully incorporated into PG. Interestingly, a small right shift of the diffraction peaks is observed in the GC compared to the SrF2 powders. This is attributed to a decrease in the interplanar spacing (d) according to the Bragg equation, which implies ion doping in SrF2 crystals. Notably, the diffraction peaks of GCs are sharper than the peaks for the SrF2 powders, indicating that growth of crystalline SrF2 occurred during the preparation process.

Figure 1
figure 1

XRD pattern of Er3+-doped PG, Er3+-doped GC and SrF2 powders.

Full size image

SEM images showing the morphology of the SrF2 powders and the fractured surface of the Er3+-doped GC containing SrF2 are shown in Fig. 2(a,b), respectively. As shown in Fig. 2(a), the average size of the SrF2 powders is ~300 nm. In Fig. 2(b), we can see that the fractured surface of the GC is rough and large amounts of isolated particles are embedded in the glass matrix. No pores are observed in the magnified SEM image (Fig. 2(b), inset, lower left), which indicates that PG has the ability to remove pores. Pores can be easily removed by dissolution in the liquid glass phase and outward diffusion to the environment because of the low viscosity of FP glass. Based on the XRD pattern shown in Fig. 1, these isolated particles are considered as SrF2 crystals. To confirm this conclusion, elemental mapping was used to discern these isolated particles, as presented in Fig. 2(b). The analysis shows that Sr, F, Ca and Er signals are detected in the isolated particle of GC, whereas the presence of Sr and F is stronger than other elements. This result indicates that the isolated particle is SrF2 and that Er3+ ions together with Ca2+ ions are incorporated into the SrF2 crystal lattice. The extra Ca is attributed to ion exchange during liquid-phase sintering. This observation supports the abovementioned doping assumption. The substitution of the larger Sr2+ (1.12 Å radius) with the smaller Ca2+ (1.00 Å radius) and Er3+ (1.00 Å radius) during the preparation process shrinks the crystal lattice. As a result, the interplanar spacing of SrF2 decreases, as shown in Fig. 1. Some information on the quality of the isolated particles, such as size and distribution, can be obtained by direct imaging using the transmission microscope. Fig. 2 (d) displays the transmission microscope micrograph of the interior of the GC. The nearly spherical particles observed in the transmission micrograph correspond to SrF2, as identified by the XRD and EDS results. It can be seen that many SrF2 particles that are ~15 μm in size are randomly distributed in the glass matrix. For comparison, the transmission microscope micrograph of PG is shown in Fig. 2(c). There is no SrF2 phase and only the glass phase is observed. Note that the size of the SrF2 particles inside the GC is a few tens of micrometers, while that of the SrF2 powders is hundreds of nanometers; this result is consistent with that of XRD pattern analysis which shows the growth of the SrF2 crystals during liquid-phase sintering.

Figure 2
figure 2

SEM image of the SrF2 powders.

(a), SEM image of the fractured surface of GC (b) with the associated Sr (blue), F (red), Ca (blue-white) and Er (green) elemental mappings in the insets. The inset in the lower left of (b) shows the magnified SEM image of the fractured surface of the GC. The transmission microscope micrograph of the interior of the PG (c) and GC (d).

Full size image

The transmittance is a very important factor in the assessment of the optical performance of transparent GCs. Based on the Rayleigh-Ganz particle scattering theory24, the grain size in a large number of GCs is limited to less than the corresponding wavelength to maintain high transparency. Generally, GCs are translucent or even opaque when the size of the grains reaches the micron level because of strong light scattering9, which indicates a poor potential for optical applications. Interestingly, a transparent Er3+-doped GC containing micron-sized SrF2 crystals was obtained by the abovementioned liquid-phase sintering route, as shown in the inset of Fig. 3, which contained crystals with diameters of approximately 15 μm. It can be seen that the average transmittance of the GC is about 75% in the visible region and more than 85% in the near-IR region. The transmittance is ~83% around 2.7 μm. The high transparency of the samples is attributed to the matching of the refractive index between the SrF2 crystals and PG. In order to examine this hypothesis, the refractive index of the SrF2 single crystal and PG was measured as 1.442 and 1.447, respectively. This result agrees with the conclusion from the in-line transmittance spectra analysis. The stretching vibration of the free [OH−] groups shows a strong absorption band at 2.5–3.0 μm, which affects the 2.7 μm emissions of Er3+ ions25,26. A small [OH–] group absorption peak (~2.85 μm) is observed in GCs, indicating that the samples possess a strong ability for water removal. Owing to the similar ionic sizes of F− and OH– ions and the isoelectronic properties, F− ions can efficiently replace OH− ions during liquid-phase sintering27.

Figure 3
figure 3

Transmittance spectrum of the Er3+-doped GC containing SrF2 (1-mm thick).

The photograph in the inset shows transparency of the GC.

Full size image

Wei6 and Fan9 have pointed out that the efficient 2.7 μm emissions of Er3+-doped GCs containing fluoride crystals is attributed to the Er3+ ions entering the fluoride crystal lattices. Thus, Er3+ ions in crystals embedded into the glass matrix play a crucial role in mid-infrared emissions. In order to confirm that the Er3+ ions are incorporated into the SrF2 lattices during the GC preparation process, the fluorescence spectra at 1.53 μm was investigated, as presented in Fig. 4. Under 980 nm laser excitation, both PG and GC show 1.53 μm emission bands attributed to the transition of Er3+:4F13/2→4I15/228. Notably, the GC exhibits stronger 1.53 μm emissions than the PG and the integrated intensity of the GC is about 2.15 times higher than that of the PG. Decay lifetimes of the Er3+:4S3/2, 2H11/2, 4F9/2 and 4I13/2 level in the prepared samples excited at 980 nm were also recorded, as shown in Table.1. Obviously, the decay lifetimes originated from the Er3+:4S3/2, 2H11/2, 4F9/2 and 4I13/2 level in the GC are much longer than that of PG. The changes observed in both the photoluminescence (PL) spectra and PL lifetimes indicate that the doped Er3+ ions are embedded in the low-phonon-energy SrF2 crystals during liquid-phase sintering.

Table 1 Decay lifetimes of the Er3+:4S3/2, 2H11/2, 4F9/2 and 4I13/2 levels in the prepared samples excited at 980 nm.
Full size table
Figure 4
figure 4

Fluorescence spectra of samples at 1.53 μm, excited at 980 nm.

Full size image

The mid-infrared emissions of the Er3+-doped PG and Er3+-doped GC containing SrF2 crystals were investigated, as presented in Fig. 5. An intense fluorescence around 2.7 μm is observed in GC and no mid-infrared emissions are generated in PG, indicating that the mid-infrared emissions result from the Er3+ ions in the low-phonon-energy SrF2 distributed in the glass matrix rather than the Er3+ ions in the glass matrix. Meanwhile, the 4I11/2 → 4I13/2 transition shows an evident Stark splitting in the GC, implying that Er3+ ions are incorporated into the SrF2 lattices and the 4I11/2 → 4I13/2 transition is influenced by the crystal field of SrF229,30. Mid-infrared fluorescence quenching of Er3+ in PG results from strong nonradiative decay due to multiphonon relaxation9, which is attributed to a large amount of high-phonon-energy phosphate in PG. To estimate the 2.7 μm emission properties, the emission cross section of GC is calculated as follows31:

Figure 5
figure 5

Mid-infrared fluorescence spectra of samples excited at 980 nm.

The inset shows the emission cross section at 2.7 μm.

Full size image

where Arad is the radiative transition probability at 2.7 μm and is the normalized line-shape function of the emission spectrum. The inset in Fig. 5 shows the emission cross section of the GC. We can see that the maximum emission cross section is more than 1.2 × 10−20 cm2, which is obviously larger than that of the Er3+-doped oxyfluoride tellurite glass (4.5 × 10−21 cm2)32, ZBLAN glass (5.7 × 10−21 cm2)23, FP glass (6.57 × 10−21 cm2)23, germanate glass (8.44 × 10−21 cm2)27, bismuthate glass (9.53 × 10−21 cm2)11 and GCs containing NaYF4 (7.05 ± 0.25 × 10−21 cm2)6. A large emission cross section indicates that the GC would be a promising candidate for IR lasers and amplifiers.

To illustrate the 2.7 μm emission behavior, the energy level diagram and energy transfer mechanism of Er3+ ions, as shown in Fig. 6, have been proposed. Electrons at the ground state absorb energy and jump into the 4I11/2 level via ground state absorption (GSA) upon excitation at 980 nm. Electrons in the excited state 4I11/2 will relax in two ways. Electrons cycle back to 4I13/2 by radiative or nonradiative relaxation. Radiative transitions of the electrons generate the 2.7 μm emissions, while the nonradiative transitions cause fluorescence quenching. Meanwhile, electrons in the 4I13/2 level can further relax to the ground state, emitting at 1.53 μm. Furthermore, the excited state absorption (ESA2: 4I13/2 + a phonon → 4F9/2) and the energy transfer up-conversion (UC) (ETU2: 4I13/2 + 4I13/2 → 4I9/2 + 4I15/2) process, which are conducive for population reduction of the 4I13/2 level and improving the 2.7 μm emissions, occur simultaneously in the electrons at the 4I13/2 level6. On the other hand, electrons in the 4I11/2 level are excited to the 4F7/2 level by the excited state absorption (ESA1: 4I11/2 + a phonon → 4F7/2)33 or energy transfer UC (ETU1: 4I11/2 + 4I11/2 →  4F7/2 + 4I15/2)34. Then, the electrons in the 4F7/2 level relax nonradiatively to the 4S3/2, 2H11/2, or 4F9/2 level and cycle back to the ground state, generating green or red emissions.

Figure 6
figure 6

Energy level diagram and energy transfer processes of Er3+ ions.

Full size image

Based on the 2.7 μm emission behaviors, the ESA2 process is beneficial for electron accumulation in the 4I13/2 level and the 4F9/2 → 4I15/2 transition. Meanwhile, the ESA2 process can improve the 2.7 μm emissions. Therefore, interactions occur between the UC emissions and the 2.7 μm emissions of the Er3+ ions. In order to understand the relationship between the mid-infrared emissions and the UC emissions of the Er3+ ions, the UC emissions spectrum of GC and PG is shown in Fig. 7. Under 980 nm laser excitation, the emission peaks at 521, 540 and 653 nm in Fig. 7 are attributed to the transition of Er3+:2H11/2 → 4I15/235,4S3/2 → 4I15/235 and 4F9/2 → 4I15/236, respectively. It can be seen that the intensity of the red light emissions is much stronger than that of the green light emissions in PG and GC, which is ascribed to the ESA2 process. It is noted that the GC exhibits a much stronger UC luminescence than the PG and the integrated intensity of the GC is about 4.5 times higher than that of the PG. Meanwhile, the degree of growth of the red light emissions is larger than the green emissions in the GC. This result indicates that the low-phonon-energy SrF2 crystals make the ESA2 process, as well as the ESA1 and EUT1 process, stronger. The UC emissions spectrum indicates that the incorporation of Er3+ into the SrF2 crystal lattices results in an enhancement of both the UC and mid-infrared emissions.

Figure 7
figure 7

UC emissions spectra of samples excited at 980 nm.

The inset shows the UC emissions at 500–600 nm.

Full size image

Conclusion

Er3+-doped transparent GCs containing micron-sized SrF2 crystals were successfully prepared using a liquid-phase sintering route. The SrF2 crystals were ~15 μm in size and were uniformly distributed throughout the fluorophosphate glass matrix. The GCs achieved average transmittances of 75% in the visible region and more than 85% in the near-IR region by matching the refractive index of the precursor glass with that of the SrF2 crystals. The EDS, PL spectra and PL lifetime results indicate that Er3+ ions were incorporated into the SrF2 micron-sized crystals. Intense 2.7 μm emissions derived from the transition of Er3+: 4I11/2 → 4I13/2 were observed upon excitation at 980 nm using a laser diode because of the incorporation of Er3+ ions into the low-phonon-energy micron-sized SrF2 crystals. The maximum emission cross-section of the Er3+:4I11/2 → 4I13/2 transition is more than 1.2 × 10−20 cm2 at 980 nm excitation, implying that there is good potential for fabricating transparent GCs containing micron-sized SrF2 crystals with good spectroscopic properties for mid-IR lasers.

Additional Information

How to cite this article: Jiang, Y. et al. Er3+-doped transparent glass ceramics containing micron-sized SrF2 crystals for 2.7 µm emissions. Sci. Rep. 6, 29873; doi: 10.1038/srep29873 (2016).