1. Introduction

Heavy metal fluoride (HMF) glasses have shown great application potential in the field of high-power fiber delivery and laser sources in the mid-infrared (MIR) spectral region due to their low phonon energy, wide infrared transmission band, high rare earth (RE) ion solubility, and high laser damage threshold [1–4]. In 1974, Lukas and Poulain reported HMF glass based on the composition of ZrF₄-BaF₂-NaF for the first time [5]. Subsequently, this research brought the development of HMF glass to a climax in the world, principally in France, the United States, and Japan. In 1981, Ohsawa et al. reported ZBLAN glass, the most stable and practical HMF glass for MIR optical fiber applications thus far [6]. ZBLAN fiber doped with RE ions has shown laser emission spanning 2.8–3.9 µm [7–9]. However, longer emission wavelengths in fluorozirconate glass are almost impossible to achieve due to the increased influence of multiphonon relaxation and the exponential increase in fiber loss beyond 3.8 µm [10]. With the increasing demand for laser sources and fiber laser delivery in the entire MIR 3–5 µm region, it is necessary to develop an HMF glass system with lower phonon energy than ZBLAN.

Fluoroindate glasses with lower mean values of phonon energy (510 cm⁻¹) than fluorozirconate glasses (550 cm⁻¹) have attracted attention in recent years [11]. Fluoroindate glasses have been developed by adding In³⁺ to fluorozirconate glasses. In³⁺ prevented reduction of Zr⁴⁺ to Zr³⁺ and can also slightly increase the resistance of fluorozirconate glasses to crystallization [12–14]. As the In³⁺ content slowly increased, it gradually developed into fluoroindate glasses. InF₃ alone cannot be vitrified, and it is considered as a glass former of the second degree, because the glasses are obtained in binary systems at ordinary cooling rates [15–18]. To date, numerous fluoroindate glass components have been reported; however, the continuous optimization of fluoroindate glass composition is still in progress due to their crystallization tendency being much greater than that of fluorozirconate glass. However, these fluoroindate glasses possess excellent optical properties, which make them the focus materials in the MIR region. The transparent window of fluoroindate glasses extends to approximately 7–8 µm in the MIR region, which is longer than that of fluorozirconate glasses (∼ 5 µm), and the theoretical fiber loss of fluoroindate glass was calculated to be lower than that of fluorozirconate glass [17,19]. This shows that fluoroindate glass fibers have potential applications in optical fiber communication. However, the loss of fluoroindate glass fiber is higher than the theoretical value due to the technical problems existing in the preparation of fluoride glass and optical fiber drawing, which requires further development and innovation of technology. Moreover, at present, fluoroindate glass is usually considered as a nonlinear medium to broaden the spectral coverage of fiber laser sources through soliton self-frequency shift, which is called SC sources [20,21]. Additionally, the high RE ion solubility of fluoroindate glass also shows its application potential in the field of bulk and fiber lasers.

This review summarizes the development and research status of fluoroindate glass, including component optimization, glass structure research, laser performance, and preparation technology for bulk glass and optical fibers. In addition, we have introduced our related progress in fluoroindate glass and optical fiber. The last part of the article looks forward to the prospects of fluoroindate glass and optical fibers.

2. Optimization of fluoroindate glass fiber compositions

In the early 1980s, the composition of fluoride glass containing InF₃ was reported for the first time [22], where a small amount of InF₃ was added to divalent fluoride glass to improve its crystallization resistance. Divalent fluoride glasses in ZnF₂-SrF₂, MnF₂-BaF₂, ZnF₂-BaF₂-SrF₂, ZnF₂-BaF₂-CdF₂, and ZnF₂-BaF₂-SrF₂-CaF₂ with weak crystallization resistance can only obtain samples of 2–3 mm. The incorporation of InF₃ resulted in stable glass samples with more than 10 mm thickness and expanded its infrared transmission range, making fluoroindate glasses a promising HMF glasses [23–26]. The simplest component of fluoroindate glass was based on the binary system InF₃-BaF₂, whose glass-forming ability was weak. Additives such as MgF₂ and GaF₃ were proposed to stabilize the basic binary system [27]. The addition of MgF₂ improved the resistance to devitrification more efficiently than that of GaF₃. Nevertheless, it should be noted that the small atomic weight of Mg resulted in an unexpected MIR transmission edge shifting towards shorter wavelengths [28]. Therefore, 5 mol% MgF₂ was considered the optimal doping, which balanced the glass stability and transmission window for fluoroindate glass applications [27].

Increasing the number of components can be conducive to improving the resistance to devitrification of glass, which was considered the “confusion principle” [29]. Consequently, SrF₂, ZnF₂, CaF₂, MgF₂, YF₃, PbF₂, and GaF₃ were used as additives to stabilize the glass [26,30,31]. Among them, the most widely studied system was InF₃-BaF₂-SrF₂-ZnF₂-MF_n (M = Na, Ca, Mg, Cd, Pb, Gd, Y, Zr), as shown in Table 1. The improved glass-forming ability of the additives was evaluated as follows: Ga > Mg > Cd > Y > Zr > Ca > Na > Gd [32]. Strikingly, the system InF₃-BaF₂-SrF₂-ZnF₂ containing 4 mol% GaF₃ or MgF₂ has good thermal stability and a smaller activation energy. Moreover, the doping of GaF₃ or MgF₂ also made it possible to produce a preform for fiber drawing. The addition of GdF₃ improved the feasibility of obtaining large bulk fluoroindate glass [33], with an optimum doping amount of approximately 4 mol%. The preform, with a size of Φ 10 mm × 120 mm, and the fiber were successfully prepared and exhibited better corrosion resistance than fluorozirconate glass [34].

Table 1. The compositions based on InF₃-BaF₂-SrF₂-ZnF₂.

View Table | View all tables in this article

Further efforts on the variations of the compositions were performed to enhance the resistance to the devitrification of glass. The RE ions were considered to be intermediate between glass progenitors and modifying cations [16]. Moreover, the chemical bonds of fluoroindate glass are weak, and the field strength of In³⁺ is close to that of RE ions, which resulted in the easy introduction of RE ions into the glass network [35]. LaF₃, CeF₃, PrF₃, NdF₃, TmF₃, ErF₃, and EuF₃ were incorporated into the fluoroindate glass [36], and their effects on the glass were investigated. It was observed that the resistance to devitrification was obviously increased when the RE ion doping level was below 3 mol%. Notably, ΔT reached 114 °C when 2 mol% NdF₃ was incorporated.

Nevertheless, superficial devitrification occurred on fluoroindate glasses, resulting in mechanical and optical defects, which affected the application of optical fibers. The composition of the fluoroindate glasses contained a considerable amount of ThF₄, which seemed to be more stable. The ThF₄ contained system with a composition of BaF₂-InF₃-ZnF₂-ThF₄ and BaF₂-InF₃-ZnF₂-YbF₃-ThF₄ was observed [19]. However, the natural radioactivity of ThF₄ made it unsuitable for industrial applications.

Maze et al. [39] reported an optimization method for the composition of fluoroindate glasses. In a stable fluoroindate glass system, the composition of InF₃, ZnF₂, and MF₂ should not be less than 70% by weight, where M can be one of the Ba, Sr, Ca, or Pb. CdF₂ and MgF₂ can be introduced into the glass as stabilizers to improve the resistance to devitrification, with a content of approximately 20% by weight. Furthermore, InF₃ and ZnF₂ can be partially replaced by GaF₃ and MnF₂, respectively, which can be carried out separately or simultaneously. Consequently, the crystallization resistance of fluoroindate glass can be effectively improved and the corresponding fiber attenuation is 0.06 dB/m @ 2.7 µm. Moreover, it should be noted that the addition of Mn can reduce the transmittance in the near-ultraviolet region, and the glass samples turn yellowish.

From the application point of view of fluoroindate glass, the characteristics of cations should be considered in the design of components. Ga³⁺ ions, leading to an increase in the phonon energy of the glass system, should preferably be avoided when pursuing a wider infrared transmission window. To control the refractive index of fluoroindate glass for the appropriate core/cladding combinations of fibers, the polarizability of cations should be taken into consideration. The refractive index of glass depended on the polarizability of the cations in the composition. For example, the refractive index of glass decreased with indium substituted by magnesium, a lighter and lower polarizable element than In [27]. To reduce the phonon energy of the glass matrix, highly concentrated trivalent fluorides were not recommended. This is because the trivalent fluorides would increase the fundamental vibration energy. Adam et al. [40] attempted to reduce the InF₃ and GaF₃ contents to reduce the system phonon energy. It was observed that the best balance between glass stability and low phonon energy was achieved in fluoroindate glass with 10 mol% trivalent fluorides (InF₃ + GaF₃).

The components of the fluoroindate glass system are complex and diverse, which is different from that of fluorozirconate glass. Consequently, the composition design of fluoroindate glass should balance the stability and functionality according to the demand, especially the functional optical fiber.

3. Structure of fluoroindate glass

The lack of fixed structural units has restricted the research on the structure of fluoride glasses. For example, the average coordination (CR) number of Zr in fluorozirconate glass is not a definite value. This is likely because of the ZrF₄ crystal, which shows Zr with CR of 6, 7, and 8. This also indicates that there are different structural units in the glass network. Reports on the structure of fluoroindate glass are very scarce. However, clearly revealing the glass structure is important in terms of optimizing the composition of the fluoroindate glass and designing the glass performance. In the 1990s, there was an upsurge in research on the structure of fluoroindate glass. Advanced methods such as spectroscopic and diffraction techniques revealed the structural information of glass materials. The majority of spectroscopic techniques include Raman scattering, infrared (IR) spectroscopy, and X-ray photoelectron spectroscopy (XPS). Diffraction techniques include nuclear magnetic resonance (NMR), X-ray and neutron diffraction, and extended X-ray absorption fine spectroscopy (EXAFS).

Notably, the coordination environment of atoms in glass is considered to be similar to the corresponding crystalline materials [41]. Kawamoto et al. [42] investigated the F1s binding energies (FBE), measured by XPS, of fluoroindate glass based on the compositions of 45InF₃-40BaF₂-15ErF₃, and InF₃ crystals, which were 684.6 eV and 685 eV, respectively. This indicated that the FBE of InF₃ crystals was related to the electrostatic interaction between F- and the surrounding cations in the glass. Meanwhile, it was observed that the FBE of the fluoroindate glass was approximately equal to the calculated sum of the FBE of the crystals constituting the glass in a molar ratio, which suggested that the FBE of fluoroindate glass possessed additive properties. The structure of InF₃ crystals was considered a framework consisting of [InF₆]^3- octahedra, and the octahedra were connected by their vertices through the bridging F atoms [43]. Similarly, Mastelaro investigated the structure of binary glass based on InF₃-BaF₂ and InF₃-SrF₂ through EXAFS and Raman spectroscopy [44]. It was found that the glass structure was built by [InF₆]^3- octahedra, and the surrounding cations (Ba²⁺ and Sr²⁺) did not participate in the construction of the network structure, as shown in Fig. 1. The EXAFS spectra of glass were investigated, and it was observed that the first neighbor’s number (N_{In F}) of In atoms (N_{In F} =6) and the bond length (R _{In F}) of In-F (R _{In F} =0.205 nm) rarely changed with the different contents of divalent cations (Ba²⁺ and Sr²⁺) doping in glass. Notably, the values were the same as those for the InF₃ crystal.

 

Fig. 1. Polarized and depolarized Raman spectra for three glasses in binary system [44].

Download Full Size | PDF

Almeida et al. [11] studied a series of fluoroindate glasses including binary and multicomponent glass by IR and Raman spectroscopy. In binary glass based on 60InF₃-40BaF₂, the vibrational modes were identified as the vibrational modes of the [InF₆]^3- octahedra without any vibrations of Ba²⁺, which was considered the network modifying cations. In fact, the vibrational modes on the IR and Raman spectroscopy were assigned to the bending or stretching vibrations of non-bridge F atoms. Given that the atomic ratio of In to F was 4.33, which indicated that each [InF₆]^3- octahedron shares approximately three F atoms with the surrounding octahedra. This also showed that the octahedron in fluoride glass structure is usually distorted, which differed from the InF₃ crystal [17]. While in the multicomponent glass based on the 20InF₃-15BaF₂-20SrF₂-40ZnF₂-CdF₂ and 40InF₃-15BaF₂-20SrF₂-20ZnF₂-CdF₂, the Zn atoms presented octahedra with six coordinates, similar to ZnF₂ crystal [45]. Notably, Zn also contributed to increase the number of bridging F in glass.

Furthermore, the structure of fluoroindate glass containing other cations was investigated. Zhu et al. [46] studied 15InF₃–20GaF₃–20PbF₂–15ZnF₂–20CdF₂–10SnF₂ glass using EXAFS and XRD spectroscopy. The Ga atoms were considered to play a similar role to the In atoms in the glass structure, that is, Ga atoms presented six coordinates and were the network builders. Furthermore, Ga atoms were found to be more stable than In atoms because the In-related crystalline phase was precipitated from the glass matrix after heat treatment, while Ga-related crystalline phase was not found. Zn atoms were recognized as glass network modifiers with six coordinates, and the coordination radius increased once tetragonal ZnF₂ was precipitated from the glass structure due to the distortion of the Zn–F bond. This result suggests that Ga³⁺ is a good stabilizer for fluoroindate glass because it can participate in the construction of the glass network structure, which is well documented by Bakhvalov et al. [47] GaF₃-based glasses and fluoroindate glasses were studied using IR and NMR spectroscopy. In fact, GaF₃-based glass consisted of [GaF₆]^3- octahedra, according to the IR data. However, the glass-forming octahedrons were connected in different ways between the GaF₃-based and fluoroindate glasses. That is, the [GaF₆]^3- octahedra were linked by three edges, while the [InF₆]^3- octahedra were connected by two edges and two vertices, as shown in Fig. 2, which is consistent with the results of previous studies [11]. Notably, the [InF₆]^3- octahedrally linked pathway depended on the nature and quantity of additives.

 

Fig. 2. Structure nets of GaF₃-based (a) and fluoroindate (b) glasses [47].

Download Full Size | PDF

When the alkaline fluoride is introduced to the fluoroindate glass, one subsystem unit changes with temperature while irrelevant toward the other. The lack of holistic change indicated that alkaline fluoride cations were assigned to different positions in the fluoroindate glass structural net. Similarly, Ignat'eva et al. [48] studied the structure of glass based on InF₃-BaF₂-AlF₃-PbF₂-LiF by IR spectroscopy. Moreover, it was observed that the addition of Li⁺ reduced the connectivity of the octahedron in fluoroindate glass. Another argument is that there are different glass formers in the fluoroindate glass structure that lead to different subsystems. Similarly, a change in the degree of connectivity of the [InF₆]^3- octahedron occurred in the glass network when BiF₃ was introduced into the glass [49]. That is, the degree of connectivity decreased with the doping of BiF₃ because of the appearance of Bi-F polyhedra in the glass structure [50]. As a consequence, T_g of the glass decreased with the doping of BiF₃ due to the change in the diffusion motion of F⁻, as revealed by 19F NMR spectroscopy [51]. Moreover, it was observed that Bi³⁺ as the network modifier enhanced the fluorescence level of RE ion-doped fluoroindate glass.

Azkargorta et al. [52] studied the structure of Nd³⁺-doped fluoroindate glass by investigating its optical properties, because the optical properties closely depended on the coordination environment and the bonding at the RE ion site [53]. The ⁴I_9/2→⁴F_3/2 transition of Nd³⁺ emitting two bands of fluorescence at 1049.6 and 1050.6 nm suggested the Nd³⁺ assigned in different site in glass structure. Kochanowicz et al. [54] studied the structure of rare-earth Er/Ho co-doped fluoroindate glass by IR spectroscopy, as shown in Fig. 3. The addition of Er/Ho enhances the electron cloud of F⁻ around the [InF₆]^3- octahedra, which led to the IR shift to a low wavenumber [55,56]. Therefore, Er/Ho was recognized as a glass network modifier that can influence the [InF₆]^3- octahedra depolymerization.

 

Fig. 3. MIR spectrums of (a) undoped glass and (b) Er³⁺/Ho³⁺ co-doped glass [54].

Download Full Size | PDF

Numerous studies have focused on the effect of additives on the structure of fluoroindate glass, while little attention has been given to the effect of temperature. Recently, the structural evolution of fluoroindate glass at high temperatures by Raman spectroscopy was studied by our team [57], as shown in Fig. 4. The high-frequency and low-frequency peaks were assigned to the symmetric stretching of non-bridge F atoms without network cation motion and the bending vibration of non-bridge F atoms, respectively [11]. The Raman peaks were found to move to a lower wavenumber as the temperature increased, which is due to the thermal expansion of the In-F bond at high temperatures. It is noteworthy that the intensity of the stretching vibration of non-bridge F atoms became weak, but the bending vibration became stronger as the temperature increased. This indicated that the number of nonbridging F atoms increased at high temperatures. These results suggested the dissociation of [InF₆]^3- octahedrons in the glass structure at high temperatures. Further, the Raman peak returned to its original state after cooling to room temperature, which showed the change in the glass structure at temperatures below Tg was reversible.

 

Fig. 4. Raman spectroscopy under different temperatures [57].

Download Full Size | PDF

Thus far, researchers have agreed that fluoroindate glasses are recognized as octahedral glasses based on the CR numbers of the glass structure former cations. Additives affect the connection of [InF₆]^3- octahedrons; nevertheless, this depends on the additive nature and amount. Therefore, to optimize their composition, information on these glass structures is essential.

4. Preparation technology of fluoroindate bulk glass and optical fiber

Considering the preparation cost and market scale of fluoride bulk glass, the development of fluoindinate glass has mainly been focused on fiber applications. Compared with fluorine zirconium fiber, the transmission range of fluoindinate fiber can span a wider IR range and has a lower transmission fiber loss. Indeed, the infrared transmission cut-off edge of fluoindinate glass can reach 7.5 µm corresponding to a thickness of 20 mm [58], which is wider than that of fluorozirconate glass, as shown in Fig. 5. The theoretical transmission loss of the fluoride fiber has been estimated to be approximately 10⁻³ dB/km at 2.6–3.5 µm, which is 100 times lower than that of silica fiber [59]. However, to date, the loss value of the prepared optical fiber is far from the theoretical value, which is closely related to the manufacturing technique. This is the current bottleneck of fluoride fibers. In 2018, SpaceX manufactured a fluoride optical fiber under a microgravity environment in space, which contributes to reducing ion migration, reducing the crystallization tendency. The prepared fiber loss was not available, but the mechanical properties of the fibers were effectively enhanced. This further demonstrates the difficulty in manufacturing low-loss fluoride fibers.

 

Fig. 5. The transmittance of ZBLAN and fluoindinate glass [58].

Download Full Size | PDF

Several studies have focused on the preparation of low-loss fluorinated glass fibers. Yoshiki Nishida developed a fluoindinate glass fiber based on InF₃-GaF₃-PbF₂-ZnF₂-YF₃-LaF₃, which presented a thermal stability similar to that of fluorozirconate glass. This glass system was fabricated to optical fiber with loss of 43 dB/km @ 3.36 µm [60]. Saad reported a new fluoindinate optical fiber with very high mechanical strength, with a transmission loss of 45 dB/km @ 2–4 µm [58]. YJestin investigated a new multicomponent fluoindinate glass based on the composition of InF₃-PbF₂-GaF₃ with a loss of 850 dB/m @ 1.3 µm [61]. Till now, the lowest loss record of the fluoindinate glass fiber manufactured in the laboratory is 0.002 dB/m at 3.6 µm, and the loss of the commercial fluoindinate glass fiber (Le Verre Fluoré) is 0.01 dB/m. The realization of these loss levels is due to the improvement in modern raw material purification technology and optical fiber preparation technology. This is because the loss in the fluoroindate glass fiber was mainly caused by Rayleigh scattering and absorption loss. The obvious crystallization tendency of fluoride glass leads to an increase in the number of grain scattering centers. Moreover, impurities in raw materials cause Rayleigh scattering. The absorption loss mainly originates from transition metals, RE ions, and hydroxyl groups. Therefore, modern technology for fabricating low-loss fluoindinate fiber focuses on the purification of raw materials and the removal of impurities and hydroxyl in glass melting. Here, the glass melting process and fabrication technology of optical fiber are discussed.

InF₃ is extremely sensitive to moisture, especially at high temperatures; therefore, it tends to be hydrolyzed to form hydroxides or oxides. Usually, transparent fluoroindate glass cannot be obtained without atmosphere protection during the melting process, which results in an opaque yellow bulk material. This characteristic makes it difficult to produce transparent fluoroindate glasses. Simultaneously, it is also unfavorable for the preparation of low-loss InF3 glass fibers. In modern technology, reactive atmosphere processing (RAP) is widely used in the preparation of fluoindinate glass, which uses reactive gases such as CCl₄, SF₆, and NF₃ to react with hydroxyl in glass melt to reduce the OH⁻ content. However, the introduction of Cl⁻ has been proven to be harmful to fluoride glass [62]. Therefore, the use of an active atmosphere containing Cl⁻ should be avoided as much as possible.

Moreover, the oxide in the raw material produced during the melting process also significantly affects the properties of fluoindinate glass. The ammonium bifluoride method is a very effective method for fluorinating oxide impurities. Ammonium bifluoride decomposes to produce NH₃ and HF, the latter of which can lead to the oxide transition to fluoride and contribute to providing fluoride atmosphere protection for glass melt. However, it should be noted that the reactions between ammonium bifluoride and fluoindinate glass batches are better at low temperatures, as the ammonium bifluoride tend to hydrolyze under high temperatures. The recommended reaction temperatures were in the range 200-500 °C.

Bekafang has extensively worked on optimizing the melting conditions of fluoindinate glass. Ammonium bifluoride was added to the glass batches under the protection of N₂ atmosphere, which can effectively reduce the interference of oxygen in the air. The reactions were performed at 235 °C and then heated to a melting temperature of 900 °C. The absorption of hydroxyl of obtained fluoindinate glass samples was significantly reduced to less than 0.15 cm⁻¹ @ 2.9 µm [63]. Furthermore, the pretreatment time and temperature of ammonium bifluoride were optimized. Additionally, SF₆ was introduced as a reactive atmosphere to investigate its effects on dehydration. The conditions for the fabrication of the fluoindinate glass are presented in Table 2. Pretreatment with ammonium bifluoride is a key process. Notably, unmelted particles were found in the glass sample after pretreatment at 235 °C. This is presumably because ammonium bifluoride does not completely react at low temperatures. Increasing the pretreatment temperature, such as 450 °C and 500 °C, can effectively remove residual ammonium bifluoride [64,65]. Moreover, SF₆ was better introduced at a low temperature (700 °C) because it can erode the platinum crucible at high temperatures. In contrast, it was found that the thermal stability of fluoindinate glass was better after pretreatment with ammonium bifluoride.

Table 2. The conditions for fabrication of fluoindinate glass [63,65,66].

View Table | View all tables in this article

Jha et al. optimized RAP for melting fluoindinate glass. The melt conditions are listed in Table 3. An active atmosphere was used to pretreat the glass batch at low or high temperatures for a duration instead of being induced directly into the melt. Then, the treated powders were transferred into a glove box protected by a dry N₂ atmosphere and heated to the melting temperature for melting. Notably, the melting process should preferably be completed within 45 min, which could minimize the erosion of moisture generated by RAP and dwell inside the glove box. SF₆ is considered an effective active atmosphere for reducing the oxygen concentration at high temperatures because it can undergo redox reactions with oxide or oxyfluoride to form the corresponding fluoride. However, for the use of HF, it produces H₂O during the deoxidation process, which is harmful to the fluoindinate glass. Usually, a mixture of HF and SF₆ can produce high-quality fluoindinate glass. Sample No.6 was proven to be the most stable glass, which achieved fibers with a loss of 12 dB/m at 500 nm and 5 dB/m at 1300 nm of 1–2 m in length. However, ammonium hydrogen is considered to be less effective in removing oxygen because it decomposes above 250 °C. This leads to incomplete fluorination, thereby resulting in crystallization.

Table 3. The conditions of RAP for fluoindinate glass [67].

View Table | View all tables in this article

As high-resistance devitrification and transparent fluoindinate glass is difficult to fabricate, ammonium bifluoride or RAP is necessary. Their effects depend on the route of melting of the fluoindinate glass. It is worth mentioning that the reaction atmosphere is prone to corrode platinum crucibles at high temperatures; therefore, it is necessary to select corrosion-resistant crucibles, such as glassy carbon crucibles. The surface in contact with the vitreous carbon crucible will have a thin layer of carbon or fluorocarbon, which can be removed by grinding, hydrolysis, or ion reactions. At the same time, it can avoid precious metal pollution caused by the platinum crucible. Ammonium bifluoride should be avoided in a glove box, whose decomposition will corrode the glove box.

Fluoindinate glass fibers are fabricated using the traditional preform technique, which is related to their low viscosity. This process makes it easier to control the thickness ratio and concentricity of the core and cladding. For unclad fibers, preforms can be prepared using casting and extrusion methods. Notably, the casting method inevitably generates bubbles, which will become a new scattering center in the optical fiber. Extrusion can avoid the generation of bubbles and increase the surface finish of the preform. In contrast, the loss of fiber fabricated by the casting method was 42.3 dB/m, while the loss of the extrusion method was 2 dB/m [63]. The two fibers obtained were based on the same batch of fluoindinate glass. The reduction in fiber loss of the extrusion method was due to the preforms with a low content of bubbles and better surface finish. However, the extrusion method has strict requirements for molds. Considering the friction between the billet and the mold, graphite is widely used as a mold material for fluoride glass because of its smooth surface [63,68]. It can allow fluoride glass to be extruded at the maximum viscosity, which will greatly reduce the tendency of fluoride glass to crystallize during extrusion.

For the fluoindinate glass fiber with a complete core/cladding structure, the continuous casting or rotary casting method is usually used, as shown in Fig. 6. In comparison, the rotary casting method allows the fabrication of long preforms with good concentricity and uniform diameter, and to control the size of the core and cladding well. However, these two methods inevitably produce bubbles, which should be improved from the perspective of the process.

 

Fig. 6. Continuous casting (a) and rotary casting (b) methods for performs fabrication.

Download Full Size | PDF

5. Spectral properties of fluoindinate glass and fiber

5.1 Upconversion of fluoindinate glass

Fluoindinate glass has a high solubility of RE ions due to its weak bond strength, which makes it an ideal host material. In recent years, much attention has been paid to fluoindinate glasses doped with RE ions due to their potential applications in upconversion (UC) lasers, MIR lasers, fiber optical amplifiers, and SC [69–72]. In particular, the UC properties of fluoindinate glasses have been widely investigated. The UC is a process that converts low-energy photons to high-energy photons, which can emit visible light after excitation by infrared lasers [73,74]. Consequently, the UC properties of fluoindinate glasses show great potential for use in laser displays, infrared detection, and thermal sensors [75,76]. Numerous studies on the UC of fluoindinate glasses doped with Er³⁺, Ho³⁺, Nd³⁺, Pr³⁺, Tm³⁺, Yb³⁺/Ho³⁺, Yb³⁺/Tm³⁺, and Yb³⁺/Tb³⁺ have been reported. Here, we discuss the UC properties of RE ion-doped fluoindinate glasses.

Araújo and coworkers have conducted a series of studies on the UC of Er³⁺-doped fluoindinate glasses. Their work demonstrated that Er³⁺-doped fluoindinate glasses can emit multi-wavelength fluorescence after excitation by an IR laser diode (1.48 µm), which has aroused great interest [70,77,78]. The related UC processes are complicated because multiple processes can result in a population of excited states of Er³⁺ after excitation by 1.48 µm. The dominant process was energy transfer (ET) rather than excited state absorption (ESA), because the large concentration of Er in fluoindinate glass samples increased the interactions between Er-Er, and the ESA process was less likely to occur owing to the weak laser intensity [79]. The ET process began with the ground state absorption ⁴I_15/2→⁴I_13/2, after which the ET process occurred between two excited ions at ⁴I_13/2, which excited one ion to the higher energy level ⁴I_9/2, and then two other continuous transfer processes took the ions to the energy levels ³H_9/2 and ³H_11/2. Subsequently, radiative transitions to the low-energy level resulted in visible and infrared emissions. This multiple-wavelength emission is due to the low multiphonon decay rates of the host materials, which is due to the low phonon energy of fluoindinate glass. Notably, the UC emissions were observed to be temperature-dependent. The relative intensity between the two emissions at 530 nm and 550 nm changed with temperature, which reveals the efficient thermal coupling between the excited states ²H_11/2 and ⁴S_3/2. This behavior was considered more efficient than other fluoride glasses, suggesting that Er³⁺-doped fluoroindate glasses have potential applications in temperature sensors [78].

The UC emissions depended on the excitation wavelength. Red, green, and near-infrared emissions were observed in Er³⁺-doped fluoroindate glass after pumped by a 1550 nm laser diode, as shown in Fig. 7(a) [80]. These emission intensities showed a strong Er³⁺ concentration dependence, following a linear enhancement rule until 7 mol% Er³⁺. The mechanisms of these three emissions were considered due to the ESA and energy transfer upconversion (ETU) processes, and the related energy level diagram of Er³⁺ is presented in Fig. 7(b). Notably, Er³⁺-doped fluoroindate glass was verified to be applied in Si solar cells based on the UC properties, which can provide a current of 200 µA to the solar cell after excitation at 1550 nm with a laser power of 22 mW.

 

Fig. 7. (a) UC emissions of Er³⁺ doped fluoroindate glass pumped at 1550 nm and the pictures of the samples. (b) Possible energy transitions scheme for Er³⁺ in the fluoroindate glass after excited by 1550 nm [80].

Download Full Size | PDF

Blue, red, and green emissions could be obtained in Er³⁺-doped fluoroindate glasses when excited at 790 nm [81]. The UC processes were proved to be Er-concentrated, dependent on transient measurements, which were dominated by ET processes when the doping level was high, and for lightly doped, the dominant processes were ESA and ET. Notably, the blue emission of UC in Er³⁺-doped fluoroindate glasses is an attractive application in short-wavelength lasers. However, the efficiency of UC needs to be considered in the development of laser devices. Although a high UC efficiency was obtained in Er³⁺ single-doped fluoroindate glass [77], the Tm-Er co-doped was aimed at further improving the UC efficiency. A 60-fold enhancement of the UC intensity at ∼ 670 nm was observed in the Tm-Er co-doped system after excitation at 790 nm, which is due to the strong cross-relaxation processes between the two RE ions [82].

UC emissions at shorter wavelengths, such as blue and ultraviolet light, have attracted more attention. Tm³⁺ with stable excited levels is considered an ideal RE for obtaining blue and ultraviolet UC fluorescence [74]. In particular, intense ultraviolet (360 nm) and blue (450 nm) UC emissions were observed in Tm³⁺-doped fluoroindate glass after excitation at 630 nm, which was stronger than that of other fluoride glasses [83]. This was because the UC emission intensity depended largely on the maximum phonon energy of the host material. Red, blue, and ultraviolet UC emissions were observed in Tm³⁺-doped fluoroindate glass after excitation by a pulsed laser at 650 nm [84]. Meanwhile, the UC processes under different Tm³⁺ doping contents were investigated, and it was concluded that the dipole-dipole interaction of Tm³⁺ led to the ET process. The energy transition processes discussed by the Inokuti-Hirayama procedure contributed to determining the critical distance between Tm³⁺ for different doping levels and suggested that the Tm³⁺ had an uneven distribution in the fluoroindate glass [85]. Interestingly, the UC emission intensity of blue light was observed to be enhanced by co-doping with Yb³⁺ in fluoroindate glass [86,87], and it increased with the Yb³⁺ doping content. It was found that the UC processes involved two photons because the UC intensity presented a quadratic relationship with the pumped power at 796 nm, indicating that the dominant process was ET. This is due to the effective cross-relaxation processes between Yb³⁺ and Tm³⁺; hence, the UC efficiency was expected to increase beyond this limit.

In addition, the UC emission of blue light can also be obtained in Pr³⁺-doped fluoroindate glass [88–90]. An efficient blue emission at 480 nm was observed after excitation at 588 nm, which corresponds to the transition ³P₀→³H₄. It should be noted that the intensity of blue emission at 485 nm enhanced 20-fold times when the fluoroindate glass sample was heated to 260 °C, which used Yb³⁺ to sensitize Pr³⁺ and was excited at 1.064 µm [89]. The blue enhancement was due to the absorption cross-section, and the acceptor excited-state absorption was temperature dependent. Interestingly, a similar phenomenon was observed in Nd³⁺-doped fluoroindate glass. The UC emission in the near-infrared region of 750 nm excited at 866 nm was enhanced by approximately 40 times when the sample was heated from 298 to 498 K [91]. This phenomenon indicates that the heating process for the RE-doped fluoroindate glass can contribute to an increase in the UC laser output power.

While the Yb³⁺-Pr³⁺ co-doped fluoroindate glass was excited at 976 nm, a multi-wavelength, blue to red, fluorescence was observed [90]. The emission intensity was found to be Yb³⁺ concentration-dependent, showing a positive correlation. By comparison, no visible emissions were recorded in the single-doped samples, suggesting a strong ET between Yb³⁺ and Pr³⁺. The excited Yb³⁺ ions transfer their energy from ²F_5/2 to ¹G₄ of Pr³⁺; further, Pr³⁺ absorbed an anther pumper ion to excite the ³P₀ level for the ESA process. Finally, the radiative transitions from ³P₀ to low-energy levels emit visible light at different wavelengths.

In addition, Nd³⁺ has been widely doped in various optical materials to investigate its UC properties due to its attractive emission in the ultraviolet region [92–94]. Nd³⁺-doped fluoroindate glass-generated UC emissions containing ultraviolet radiation have been reported. A multiple-wavelength ultraviolet region spanning 350–450 nm was obtained in Nd³⁺-doped fluoroindate glasses excited by a dye laser, which provided the excitation energy for the transition ⁴I_9/2→²G_5/2 / ²G_7/2 [95]. In particular, the emissions at 354 nm and 382 nm were considered the result of the ET among Nd³⁺ ions, which was concluded from the temporal behavior. This is because of the obvious rise time in the decay curve, which is a typical characteristic of the ET process [96].

A random laser emitting ultraviolet light at 381 nm was obtained in Nd³⁺-doped fluoroindate glass after excitation at 575 nm [97]. The UC emission was assigned to ET between Nd³⁺ ions at the ²G_7/2 energy level, which corresponded to the cross-relaxation transition ²G_7/2 +²G_7/2 →⁴I_13/2+ ⁴D_7/2. The process was considered to be resonant and efficient. The rates of cross-relaxation transition could be recorded from the rise time of the 381 nm signal. This is closely related to the maximum phonon energy of the fluoroindate glass and the energy gap between the two excited levels. Accordingly, the 381 nm emission was assigned to the transition ⁴D_3/2→⁴I_9/2 [98], after estimation based on the energy gap law, which was because the excited ions on ⁴D_7/2 undergo non-radiative transition relaxation to ⁴D_3/2. The results showed that Nd³⁺-doped fluoroindate glass presented the potential application of an ultraviolet laser.

Meanwhile, the UC emissions spanning blue to green light can be obtained in Ho³⁺ excited by visible and infrared lasers, and the efficiencies were observed to be closely related to the maximum phonon energy of the host materials [99,100]. The 550, 490, 425, and 395 nm emissions corresponding to the transitions of ⁵S₂→⁵F₈, ⁵F₃→⁵I₈, ³K₇ / ⁵G₄→⁵I₈, and ⁵G₅→⁵I₈, respectively, were observed in Ho³⁺ doped fluoroindate glass excited at 640 nm [101]. Notably, there were other weak emissions such as 395 nm and 425 nm, which were the first times obtained in Ho3+-doped glass hosts due to the low phonon energy of fluoroindate glass [102]. Moreover, the blue and green emissions presented a rise time in the decay curve, indicating that the ET process dominated the UC emissions, and they were both Ho³⁺ concentration-dependent. While the Ho³⁺-doped fluoroindate glass was excited by 747 nm, blue, red, and intensive green emissions were obtained. Unlike discussed above, the intensive green emission was dominated by the ESA process, and it exhibited an avalanche-like behavior [103].

5.2 MIR spectroscopic of fluoroindate glass and fiber

In addition, it is worth discussing the MIR spectroscopic properties of fluoroindate glass due to its low phonon energy and wide infrared transparency window. The fluoroindate glass has been recognized as a competitive MIR photonic material, especially in the region 2.3–5 µm. More importantly, the laser sources in this special region present huge application potential in photoelectric countermeasures and remote sensing [104,105]. Ho³⁺, Dy³⁺, and Er³⁺ are suitable RE ions for MIR emission and have been widely investigated in fluoroindate glass.

Ho³⁺ ions possess attractive MIR laser transition of ⁵I₅→⁵I₆, and ⁵I₆→⁵I₇, corresponding to the emission at 3.9 and 2.8 µm, respectively. In 2018, lasing at 2.88 µm was achieved in Ho³⁺-doped fluoroindate glass fibers excited at 1120 nm, which is the first time that MIR lasing was obtained in fluoroindate glass fibers [106]. Moreover, a maximum output power of 54.5 mW was obtained in a 92 cm long fiber with a pumping power of 1224 mW, and the slope efficiency was 6%. Meanwhile, the length of the fibers was suggested to be sufficiently long to ensure that the pump energy was absorbed thoroughly. However, it should be noted that ineffective pumping decreases the lasing performance when the fiber is too long. Thus, finding a balance between the two is important. The dependence of the lasing performance on the fiber length is shown in Fig. 8. It was observed that the optimized length of Ho³⁺-doped fluoroindate fiber for obtaining efficient lasing at 2.88 µm was approximately 92 cm.

 

Fig. 8. (a) Laser output at 2875 nm.(b) Dependence of 2875 nm fiber laser performances on the length of Ho³⁺-doped fluoroindate glass fiber [106].

Download Full Size | PDF

However, the ∼2.8 µm emission from Ho³⁺ is limited by the lack of commercial high-power lasers matching the transition energy level to some extent [107]. Thus, sensitization by Yb/Er ions is considered a promising way to obtain an efficient 2.8 µm emission, which can excite the ⁵I₆ energy level by ET from Yb/Er [108,109]. In addition, the ET efficiency between Yb and Er was up to 61% in 0.8YbF₃/1.6HoF₃ co-doped fluoroindate glasses, and the emission intensity at 2.8 µm was 2.6 times stronger than the co-doped Er³⁺/Ho³⁺ [110,111]. This suggests that Yb³⁺ is a suitable sensitization choice for Ho³⁺ for emitting 2.8 µm in fluoroindate glasses.

The longer wavelength emission at 3.9 µm of Ho³⁺, a real attractive band for applications in space communication and new-generation photoelectric countermeasure, nevertheless, remains self-terminated. Moreover, the intrinsic emitting efficiency of the transition for the 3.9 µm was 0.33% in Ho³⁺ doped fluoroindate glasses [112]. This is mainly due to the multi-phonon relaxation, leading to a longer lifetime of the energy level of ⁵I₆ than ⁵I₅. In fact, the decay lifetimes containing radiative and non-radiative processes are 6.2 ms, and 135 µs in fluoroindate glasses, respectively. Notably, the lifetime of ⁵I₆ in Ho³⁺-doped fluorozirconate glasses is calculated to be 43 µs [113], and the shorter lifetime is mainly due to the higher maximum phonon energy of fluorozirconate glasses than that of fluoroindate glasses [114]. Studies on the basic spectroscopic parameters indicated that the emission of 3.9 µm was limited by the ETU, which was suggested to be enhanced by high-level doped Ho³⁺. In 2015, lasing at 3.9 µm was obtained in 10% Ho³⁺-doped bulk fluoroindate glass, which was apparently the first laser activity in fluoride glasses [104]. The 30 mm long fluoroindate glass was pumped from two end faces simultaneously by a Cr³⁺: LiSAF laser at 889 nm, as shown in Fig. 9. In addition, an output power of 7.2 mJ was obtained at a pump energy of 650 mJ, and the conversion efficiency was 1.6%. This study laid a foundation for the realization of MIR fiber lasers in fluoroindate glass systems. In 2018, the longest wavelength room-temperature fiber lasing at 3.92 µm was obtained in Ho³⁺-doped fluoroindate glass fiber [115]. The fiber attenuation was optimized to 0.2 dB/m in the range 3.4–4 µm, making it more suitable for efficient lasing at 3.92 µm. Notably, the concentration of Ho³⁺ up to 10 mol%, aimed at enhancing the ETU process, and the constant red emission originating from the ⁵F₅ to the ground state indicated an efficient ETU process. Then, the 23 cm-long fiber with double-clading was pumped at 888 nm, as shown in Fig. 10. Consequently, a maximum output power of 197 mW at 3.92 µm was obtained with a slope efficiency of 10.2%.

 

Fig. 9. Schematic of the laser setup. The insert picture is a fluoroindate bulk glass with Ho³⁺ doped [104].

Download Full Size | PDF

 

Fig. 10. (a) The fluoroindate glass fiber attenuation. (b) Schematic of the laser setup [115].

Download Full Size | PDF

However, considering the low absorption coefficient of ⁵I₅, co-doped sensitized ions to effectively enhance the pumping absorption efficiency via ET is necessary. Notably, an intense emission around 3.9 µm was obtained in Nd/Ho co-doped fluoroindate glass by pumping at 808 nm [116]. The excited energy level ²H_9/2 of Nd³⁺ undergoes nonradiative relaxation to ⁴F_3/2, then populating ⁵I₅ via ET. This result suggested that Nd³⁺ was a suitable sensitization choice for Ho³⁺ for 3.9 µm emission in fluoroindate glasses.

In addition, Dy³⁺ also possesses two MIR emissions of ∼3-4 µm, corresponding to the transitions ⁶H_13/2→⁶H_15/2 and ⁶H_11/2→⁶H_13/2, respectively. In 2018, MIR lasing of approximately 3 µm was obtained in Dy³⁺-doped fluoroindate fibers [10]. The 30 cm long fiber doped with Dy³⁺ at 2000 ppm was pumped at 1700 nm, achieving MIR lasing with an output power greater than 80 mW and a slope efficiency of 14%, as shown in Fig. 11. The reduction in slope efficiency compared with other Dy³⁺-doped fiber lasing was due to the attenuation of the fiber, which was determined to be approximately 2 dB/m. This high attenuation value contributed to the doped fluoroindate fiber fabrication.

 

Fig. 11. (a) The transitions for Dy³⁺ in fluoroindate glass for cascade lasing and lifetimes for the energy level. (b)Schematic of the fluoroindate glass fiber laser setup pumped at 1.7 µm. (c) The dependence of 3 µm laser output power of Dy³⁺ doped fluoroindate glass on the pump power [10].

Download Full Size | PDF

However, the ⁶H_11/2→⁶H_13/2 transition of Dy³⁺ for ∼4 µm emission is self-terminated, and the lifetimes for ⁶H_11/2 and ⁶H_13/2 are 1 µs and 500 µs, respectively, which are highly quenched in high-phonon-energy materials [117]. Notably, it is partially quenched in fluoroindate glasses due to its lower phonon energy. Hence, pulsed lasing is hopefully obtained, while a continuous lasing emission of approximately 4 µm was suggested by the cascade transition of ⁶H_13/2→⁶H_15/2 for 3 µm emission [118]. As predicted, after upgrading the resonant cavity, MIR emission beyond 4 µm was obtained in Dy³⁺-doped fluoroindate fiber when pumped at 1700 nm again. Notably, this was the longest room-temperature emission wavelength of the fluoride fibers. Interestingly, depopulating the ⁶H_13/2 laser level was considered another necessary way to obtain efficient MIR emission beyond 4 µm. For instance, introducing molecular groups or ions to reduce the lifetime of ⁶H_13/2 by ET might be a feasible method. Hydroxyl, with a large absorption strength of approximately 3 µm, was co-doped with Dy³⁺ into fluoroindate glass. The results showed that the lifetime of ⁶H_13/2 was significantly reduced to approximately 20 µs, which is shorter than that of single-Dy-doped (∼500 µs) [117]. However, it was still longer than that of the ⁶H_11/2 level, and the introduction of OH⁻ can affect the quality of glass and increase the loss of optical fiber, indicating that the method was not suitable for promoting the 4 µm emission. Here, it should be noted that the nonradiative decay of Dy³⁺ is abnormally high compared to that of other RE ions, which makes it difficult to obtain efficient emission in the MIR region. However, the nonradiative decay rates in fluoroindate fiber were estimated to be lower than that of fluorozirconate fiber, and they presented a longer lifetime of ⁶H_13/2 than fluorozirconate fiber, which is due to the lower maximum phonon energy of fluorozirconate glasses. These results have been proven by the 4 µm room-temperature emission only obtained in a fluoroindate system, and the long lifetime of ⁶H_13/2 also showed that fluoroindate systems are superior host materials to fluorozirconate systems for MIR around 3 µm fiber lasers [10].

Meanwhile, Er³⁺ doped fluoroindate glasses for MIR emissions beyond 3 µm have also been widely investigated. Notably, the ∼3 µm emission corresponds to the transition of Er³⁺: ⁴F_9/2→⁴I_9/2, and the energy gap between them is small, resulting in an emission of ∼3 µm, which is significantly dependent on the low maximum phonon energy of the host materials. In 2021, an intense emission of approximately 3 µm was obtained for Er³⁺-doped fluoroindate glass [119]. The doped glass was pumped by a 635 nm laser, which directly excited the ⁴F_9/2 energy level, resulting in the radiation transition to ⁴I_9/2, and the strongest emission was observed at the 9 mol% Er³⁺-doped level. This further indicates that fluoroindate glass can achieve a high doping level for RE ions. Furthermore, emission beyond 3 µm in the Er³⁺-doped fluoroindate glass system can also be achieved by pumping at 976 nm, which involves ground state absorption (⁴I_15/2 →⁴I_11/2) and excited state absorption (²H_11/2/⁴S_3/2→⁴F_9/2) processes, as shown in Fig. 12(a). Intense emission around 3.3 µm was observed in a 40 cm long Er³⁺ doped fluoroindate glass fiber, and the doping concentration was approximately 0.5 mol% [120]. Notably, the emission extended from around 3.1 µm to 3.85 µm was observed in the fiber when pumping by dual-wavelength of 976 nm and 1973 nm, as presented in Fig. 12(b). The central wavelengths for broad emission were ∼3.3 µm and ∼3.5 µm, respectively, which were the contributions of 1973 nm and 976 nm pumpings, respectively. Therefore, the peak of broad emission shifted from 3.3 to 3.5 µm gradually with the increasing power of 1973 nm. While a similar experimental route was performed on the fluorozirconate glass fiber, a 3.3 µm emission was not observed [121]. This indicates that the emission was closely dependent on the maximum phonon energy of the host material. Therefore, this result suggests that the fluoroindate glass fiber has huge applications in tunable MIR fiber lasers.

 

Fig. 12. (a) Energy transfer for Er³⁺ in fluoroindate glass pumped at 976 nm. (b) Emission spectra from the Er³⁺ doped fluoroindate fiber under 976 nm and 1973nm dual-wavelength pumping [120].

Download Full Size | PDF

These recent research results show that the application prospects of fluoroindate glass in the field of MIR laser light sources, in particular, the work bands can be expanded to beyond 3 µm, which can effectively fill the material vacancy in this band. However, this puts forward higher requirements for the fabrication process of fluoroindate glasses and optical fibers, which is due to the highly efficient emission in the MIR region, which can only be achieved in low-loss fluoroindate fibers.

6. Supercontinuum generation in fluoroindate fiber

Broadband laser sources, spanning 3-5 µm range, have attracted significant interest due to their potential applications in numerous fields, such as metrology, medicine, spectroscopy, and national security [122,123]. This is defined as the SC generated by the interaction between high-power pump lasing and a nonlinear optical fiber medium [124]. Numerous research findings, in particular, an increase in the average power of SCs has been obtained in fluorozirconate-based fibers. However, the spectrum of fluorozirconate-based fibers is restricted to approximately 4.2 µm [125–127], which is attributed to the multiphonon absorption edge. Therefore, SC extending to ∼5 µm can be achieved in fluoroindate-based fibers.

In 2013, the first generation of SCs in fluoroindate single-mode fiber was reported by Théberge et al. [128] The zero-dispersion wavelength of this fluoroindate fiber was 1.83 µm, and the loss was 0.1 dB/m @ 3.2 µm and 0.8 dB/m @ 5 µm. This enabled the MIR SC to be obtained in the low-loss fibers. The broadband SC from 2.7 to 4.7 µm was generated in a 9.5 m long single-mode fluoroindate fiber, as shown in Fig. 13, after pumped by an optical parametric amplifier, which generated 70 fs, 3.4 µm laser pulses. Notably, the obtained SC band was twice that of ZBLAN fibers under similar conditions. To obtain coherent SC sources, a femtosecond fiber laser with a high repetition rate and pumping near a zero-dispersion wavelength is necessary. Therefore, an approximately two-octave wide and coherent SC spanning 1.25–4.6 µm was obtained in fluoroindate fiber after pumped by a femtosecond fiber laser with 100 fs, 2 µm laser, 570 mW average power [129].

 

Fig. 13. (a) Injected energy dependent of SC generated in fluoroindate-based fiber. (b) SC generation in fluorozirconate based fiber [128].

Download Full Size | PDF

However, these reports have not achieved an SC extending beyond 5 µm, which does not benefit from the potential of fluoroindate-based glasses. In 2016, a 2.4–5.4 µm SC generated from a fluoroindate fiber was demonstrated by Jean-Christophe Gauthier [20]. Notably, the pump source was a 1.25 m long Er-doped fluorozirconate based fiber amplifier seeded by a 2.75 µm optical parametric generation source, as shown in Fig. 14. The obtained average power reached ∼10 mW with 82% of the power located in the mid-infrared region, which was closely related to the perfect fusion process between the fluorozirconate-based fiber and the fluoroindate-based fiber. In the same year, a super-flat SC covering 2–5 µm was demonstrated in a fluoroindate fiber [130]. Importantly, the pumped power and wavelength dependence of SC were investigated, and it was found that the 5-dB flatness SC was generated by pumping at 2.02 µm with 111.5 kW power, as shown in Fig. 15. Notably, there was a significant improvement in SC generation in the fluoroindate-based fibers.

 

Fig. 14. (a) Setup of the SC output in fluoroindate-based fiber. (b) Image of the fusion splice between fluorozirconate and fluoroindate based fibers. (c) SC generated in 15 m long fluoroindate-based fiber under different pump powers. [20].

Download Full Size | PDF

 

Fig. 15. (a) Setup of the SC output in fluoroindate-based fiber. (b) SC generated in the fluoroindate based fiber pumped at 2.02 µm with different energies. [130].

Download Full Size | PDF

In 2018, the first demonstration of a watt-level SC generated in fluoroindate-based fibers was presented [131]. The pump source was based on an all-fiber laser seeded by a semiconductor laser diode delivering 1.55 µm and 50 ps lasing. Further, a 2.5 W output laser was coupled to the fluoroindate-based fiber after multistage optical amplification, as shown in Fig. 16. Finally, a 1.0 W SC spanning 1–5 µm with over 2.25 µm octaves was obtained; in particular, a flatness of 6 dB was presented in the range 1.91–4.77 µm. In addition, researchers have focused on the optimization of SC pump sources to promote the development of SC. Notably, a Tm³⁺-doped fiber amplifier (TDFA) is considered an ideal SC source. An all-fiber SC source consisting of a broadband spectral spanning 2–2.7 µm TDFA system was designed by Yanget al. [132] The pump energy was injected into a 10 m long fluoroindate-based fiber through a low-loss fusion with silica fiber. Hence, a 1.35 W SC spanning 1.5–5.2 µm and a 4.06 W SC spanning 1.9–5.1 µm were obtained in this system, and the evolution of SC under different pump power was investigated, as presented in Fig. 17, which was a representative progress in the SC generated in fluoroindate-based fiber after improving the pump source system. Moreover, with a similar design system, the pump source of TDFA was seeded by a ∼1.9 µm mode-locked fiber laser, and a SC spanning 0.8–4.7 µm with the maximal output power of 11.3 W was generated in fluoroindate-based fiber. This high-power was benefitted from the low-loss splicing joint between the fluoroindate and silica fiber [133]. In addition, a watt-level >2.5 octave SC covering 0.75–5.0 µm was generated in fluoroindate-based fiber after pumping by a four-cascade of TDFA seeded by 1.9 µm, 420-ps pulses at 1-MHz repetition rate, as shown in Fig. 18 [134]. Similarly, Yehouessi et al [135]. designed an all-fiber SC pump source that can deliver multi-watt lasing at 1.9 µm via a four-cascade TDFA. Then, the 3 W SC spanning 2–4.6 µm was generated in a 60 m long fluoroindate-based fiber with a low loss of 27 dB/km @ 2 µm, and 9 dB/km @ 3.6 µm. However, the narrow SC was attributed to the fiber dispersion properties that were incompletely optimized for the pump wavelength.

 

Fig. 16. (a) Setup of the SC fiber source. (b) Spectral distribution of different amplifier [131].

Download Full Size | PDF

 

Fig. 17. (a) Evolution of SC generated in the fluoroindate-based fiber under different pump power; (b) SC under output power of 1.35 W and 4.06 W [132].

Download Full Size | PDF

 

Fig. 18. Schematic of the SC source. EYDFL, FBG, and LMA-TDF are Er³⁺-Yb³⁺ co-doped fiber laser, fiber Bragg grating, and large-mode Tm³⁺ doped fiber, separately [134].

Download Full Size | PDF

Although SC generated in fluoroindate-based fibers has made considerable progress, researchers are still optimizing the sources of SCs. In 2020, an approximately 7 W output power of SC spanning up to ∼ 4.7 µm was achieved in fluoroindate fibers after pumping by a Q-switched mode-locked Tm³⁺-doped fiber single-oscillator [136]. In addition, one of research aims was to enhance the output power and broaden the spectral range of SC sources. The SC source consisted of a 1.5 µm seed laser, fiber amplifier, and frequency-shift module, and a fluoroindate-based fiber as a nonlinear medium was designed by Yang et al. [137], as shown in Fig. 19. This SC source can obtain 11.8 W output power and spectral spanning 1.9–4.9 µm. Notably, the power of SC up to approximately 2.18 W beyond 3.8 µm was generated in the fluoroindate fiber, which was difficult for the fluorozirconate-based fiber to obtain this output power level in this spectral region.

 

Fig. 19. Schematic of the SC system. EYDFA and SMF are Er³⁺-Yb³⁺-doped fiber amplifier and single-mode fiber, separately [137].

Download Full Size | PDF

7. Conclusions

Compared with the classic fluorozirconate glass, the wide MIR transmission window and low maximum phonon energy of fluoroindate glass make it an ideal candidate for a new generation of fiber lasers and laser energy delivery media. However, till now, the composition of fluoroindate glass is still being further optimized. Based on the “confusion principle,” increasing the number of components is an effective way to improve the stability of fluoroindate glass. Meanwhile, it should be noted that additives such as trivalent fluorides are not recommended for high concentrate incorporation when pursuing a wide transmission and low-phonon material system. This is because they increase the phonon energy of the fluoroindate system. The structure of fluoroindate glass has been proven to be composed of octahedral units [InF₆]^3- connected by corners or edges. Notably, the connection method depends on the nature and content of the additives. It is worth noting that InF₃ tends to be oxidized easily at high temperatures, causing the fluoroindate glass to devitrify. Therefore, the fabrication of fluoroindate glass in a protective atmosphere is necessary. In particular, the ammonium bifluoride and RAP processes are common methods for obtaining low-hydroxyl-content fluoroindate glass.

Fluoroindate glass, which exhibits low phonon energy, benefits the rich energy levels of RE ions, which leads to excellent multispectral emission performance. Especially in the MIR region, Ho³⁺ and Dy³⁺ are suitable activators for fluoroindate glass, it can emit fluorescence around 3.9 µm and 4.3 µm. Nevertheless, the laser output of the fluoroindate glass fibers spanning 3–5 µm has brought the MIR fiber laser to a new level. To date, MIR lasers of the 100 mW level have been realized in fluoroindate glass fibers. However, compared with fluorozirconate fiber, it needs further development. Notably, the output power can be improved by optimizing the laser device. In addition, based on the non-linear nature of the fluoroindate glass fiber, it is a strong competitor of the SC laser source by injecting high-energy laser light to produce broadband emission. In particular, the MIR SC output of 100 mW, watts, and 10 W was achieved in passive fluoroindate glass fiber. It is important to note that the SC emissions have extended beyond 4 µm, which is sufficient to show the important application prospects of fluoroindate glass fiber in MIR SC laser sources.

Therefore, based on the excellent infrared properties of fluoroindate glass, it is expedient to investigate the application possibility of the bulk materials on infrared optical windows and lenses after enhancing mechanical properties via ion modification or coating. Moreover, fluoroindate fiber can be designed as a photonic crystal fiber to give full play to the advantages of MIR performance, and it is expected to be applied in the field of MIR high-power fiber delivery and supercontinuum light source.