Heavy metal fluoride glasses are the backbone of modern mid-infrared fiber technology. In the last few years, optical fibers made of fluorozirconate, fluoroaluminate, and fluoroindate glasses have reached the high level of reliability that is required for their widespread use in industrial and medical applications, which explains the rapid growth of this technology. In this tutorial review, the author, who was part of the team that discovered heavy metal fluoride glasses exactly fifty years ago, provides a summary of the glass synthesis process and elucidates the most important physical properties of heavy metal fluoride glasses. Finally, a brief overview of their main application areas is presented.

Since the first discovery of fluorozirconate glasses in 1974,1 Heavy Metal Fluoride Glasses (HMFG) have been the subject of numerous studies in the 1980s and 1990s, resulting in more than 10 000 papers, theses, and technical reports.2–4 While glasses based on beryllium fluoride had been extensively studied in the first half of the 20th century, their structure closely resembles that of fused silica.5 In contrast, HMFGs are a completely new class of non-oxide glasses that are entirely different in their chemical composition, their mechanism of glass formation, and their physical properties. For that reason, HMFGs exclude fluoroberylate glasses.

HMFGs emerged from a serendipitous discovery that was made by the Poulain brothers in 1974 at Rennes University in France. Unexpected glass formation was observed in a chemical system based on zirconium fluoride, barium fluoride, and sodium fluoride.1 From that point on, numerous new glasses were discovered in other multicomponent systems. This eventually led to the three major families of HMFGs that are widely known and in use at present: fluorozirconates, fluoroaluminates, and fluoroindates. These new vitreous materials are free of oxygen and do not contain the classical vitrifiers of industrial glasses, namely, SiO2, B2O3, Al2O3, and P2O5. They exhibit highly unique and specific optical, physical, and chemical properties that make them so attractive for many applications, and that will be discussed in this tutorial paper.

The optical transmission window of HMFGs stretches continuously from the ultraviolet (UV) part of the electromagnetic spectrum to the mid-infrared (mid-IR). While standard optical fibers made of silica glass are well suited for most applications in the visible and near infrared, fluoride glasses have emerged as the ideal materials for use in the 2–5 μm mid-IR spectral range. While previously mainly used in scientific research, fluoride fibers have recently reached a level of maturity that makes them also attractive for industrial applications. These days, fluoride optical fibers feature extremely low transmission loss values and are commercially available from several sources. In addition, fluoride fibers doped with rare-earth ions are of great interest as active mediums and are used in supercontinuum sources and new fiber lasers. The spectroscopy of rare earths in these glasses has been extensively investigated.2,6,7 The practical advantages of this vitreous matrix are a large lanthanide solubility, a low phonon energy, and a large interaction length that can be achieved in optical fibers. As a consequence, numerous new emission lines were demonstrated with fluoride glass fiber lasers.7 With the availability of powerful and affordable pump diodes, these laboratory prototypes are now being translated into commercial products and are enabling new and exciting applications, especially in the medical field.

Glass is normally obtained by cooling the initial glass forming melt until it becomes a rigid solid. A necessary pre-condition is that the liquid does not crystallize during this cooling step. Apart from BeF2, no pure fluoride exists in the vitreous state. Only the combination of several fluorides results in melts in which the crystallization rate is low enough to allow glass to be obtained by fast cooling. The first HMFG was reported in 1975 in the ZrF4–BaF2–NaF ternary system.1 In subsequent studies, other glass-forming systems based on ZrF4 (i.e., fluorozirconates) were identified. Multicomponent glasses based on the AlF3–MF2 (with M = Ca, Sr, Ba, and Pb) material group (fluoroaluminates) were also reported in the early 1980s.8 Fluoroindate glasses based on InF3 first appeared, among other more exotic glasses, in patent applications that were filed by two different research groups in 1980.9,10 However, it took several years before these glasses were extensively studied.11–13 Taking into account their chemical similarities, fluorohafnate glasses based on HfF4 are usually classified as fluorozirconates, while fluorogallate glasses based on GaF3 belong to the group of fluoroindates. Apart from these three groups (fluorozirconates, fluoroaluminates, and fluoroindates), numerous other fluoride glasses have been reported. Those are based on the transition metals V, Cr, Mn, Fe, Co, Ni, Cu,14 and also on Sc, Zn, Cd, Th, U, Sn, and Bi.15,16 However, all those various glasses have not led to practical applications to date. Typical compositions (in mol %) of the most important HMFGs are summarized in Tables IIII.

TABLE I.

Fluorozirconate glasses. Compositions in mol %.

GlassZrF4HfF4BaF2LaF3AlF3NaFLiFPbF2References
ZBLA 57  34    17  
HBLA  57 36    18  
ZBNA 52  24  20   17  
ZBLAN 53  20 20   19  
ZBLALi 53  20  20  3  
ZBLANP 48  17 20  6  
ZBLALiP 48  17  20 6  
GlassZrF4HfF4BaF2LaF3AlF3NaFLiFPbF2References
ZBLA 57  34    17  
HBLA  57 36    18  
ZBNA 52  24  20   17  
ZBLAN 53  20 20   19  
ZBLALi 53  20  20  3  
ZBLANP 48  17 20  6  
ZBLALiP 48  17  20 6  
TABLE II.

Fluoroaluminate glasses. Compositions in mol %.

GlassAlF3YF3BaF2SrF2CaF2ZnF2MgF2ThF4ZrF4NaFReferences
YABC 40 20 20 20       20  
BATY 28, 7 28, 7 20     22,6   21  
ZnBTA 20  20   40  20   22  
AFG3H 30, 2 8, 3 10, 6 13, 2 20, 2  3, 5  10, 2 3, 8 23  
ABZYZn 20 20 30   10   20  22  
AZBZnYC 20 10 20  10 20   20  22  
GlassAlF3YF3BaF2SrF2CaF2ZnF2MgF2ThF4ZrF4NaFReferences
YABC 40 20 20 20       20  
BATY 28, 7 28, 7 20     22,6   21  
ZnBTA 20  20   40  20   22  
AFG3H 30, 2 8, 3 10, 6 13, 2 20, 2  3, 5  10, 2 3, 8 23  
ABZYZn 20 20 30   10   20  22  
AZBZnYC 20 10 20  10 20   20  22  
TABLE III.

Fluoroindate glasses. Compositions in mol %.

GlassInF3GaF3BaF2SrF2CaF2ZnF2PbF2YF3CdF2NaFReferences
IYB 50  10     40   24  
IZBS 40  20 20  20     25  
IZBSC 40  15 20 20     25  
IZSBN 30  15 20  30    25  
PIGCZ 15 20    15 30  20  24  
GlassInF3GaF3BaF2SrF2CaF2ZnF2PbF2YF3CdF2NaFReferences
IYB 50  10     40   24  
IZBS 40  20 20  20     25  
IZBSC 40  15 20 20     25  
IZSBN 30  15 20  30    25  
PIGCZ 15 20    15 30  20  24  

Similar to most inorganic glasses, fluoride glasses are fabricated by cooling a clear melt fast enough to prevent crystallization. The detailed glass preparation process requires several steps, some of which are also common to oxide glasses, including the following well-known processes:

  • selecting and mixing of the starting materials,

  • melting,

  • fining,

  • casting,

  • cooling, and

  • annealing.

However, considering the unique chemical behavior of fluorides, the particular details of these processes are critical for the quality of the final glass. In comparison to oxide glasses, the processing of fluoride glasses is significantly different, and time and temperature adjustments are crucial. Most importantly, fluoride melts are very sensitive to hydrolysis and are furthermore prone to devitrification. In addition, their viscosity is very low at liquidus temperatures. The chemical reactivity of fluorides limits the possible materials that can be used for crucibles and enclosures: platinum, gold, and vitreous carbon are preferred, while silica is excluded.26 

Great care must be taken in the selection, storage, and handling of the starting materials. Fluoride glass synthesis requires a very low concentration of various impurities, with water being the most critical one. Water is a common impurity that can be found not only in many so-called anhydrous fluorides but also in reportedly non-hygroscopic compounds. Anionic oxygen that may appear as residual oxyfluoride also must be minimized. Other impurities include transition metals and paramagnetic rare earths that can induce optical absorption at various wavelengths. On the other hand, diamagnetic elements such as alkali and earth alkali metals play a less important role.

For the glass preparation, anhydrous fluorides are powdered, mixed, and finally homogenized. They are then heated to complete melting in platinum or vitreous carbon crucibles. The raw glass samples obtained by cooling usually appear grayish, and the glass must be fined to become fully transparent. The fining step consists of heating the melt at a high temperature—that is, above the liquidus temperature—in an oxidizing atmosphere. The viscosity decreases, and the melt is homogenized without stirring. Volatile species are eliminated, and reduced phases that give rise to light scattering are oxidized and dissolved. A clear and homogeneous glass is obtained upon cooling after the fining stage. Finally, the samples are annealed at a temperature close to the glass transition temperature. This annealing stage aims at removing any mechanical stresses induced by non-uniform cooling.

Optical fibers are obtained from glass preforms that are fabricated in an anhydrous atmosphere by pouring a core glass into a cladding tube. This tube may be fabricated by evacuation (“build in casting”) or by rotational casting. The overall process is described in detail in Fig. 1. For single mode fibers, a second drawing step is necessary to obtain a small diameter core.

FIG. 1.

Five steps involved in the manufacturing process of fluoride glass optical fibers.

FIG. 1.

Five steps involved in the manufacturing process of fluoride glass optical fibers.

Close modal

The most attractive feature of fluoride glasses is their large bandwidth of optical transmission that extends from the ultraviolet part of the electromagnetic spectrum to the mid-infrared, including the visible part of the spectrum. This can be seen in Fig. 2, which shows the transmission curve of 4 mm thick bulk samples made from standard fluoride glasses. The limits of transmission are determined by the chemical composition of the glasses. In the mid-IR, the multi-phonon absorption edge moves toward longer wavelengths according to the following sequence: fluoroaluminates, fluorozirconates, and fluoroindates.

FIG. 2.

Optical transmission of standard fluoride glasses: fluoroaluminates (AFG), fluorozirconates (ZFG), and fluoroindates (IFG). The transmission of pure silica (SiO2) is also shown for comparison.

FIG. 2.

Optical transmission of standard fluoride glasses: fluoroaluminates (AFG), fluorozirconates (ZFG), and fluoroindates (IFG). The transmission of pure silica (SiO2) is also shown for comparison.

Close modal
The linear refractive index n of fluoride glasses is in the medium range for typical glasses, with a value close to 1.5 in the visible. This is the result of two opposing factors: the low polarizability of fluorine anions and the large polarizability of heavy cations. The optical dispersion dn/ is expected to be low for glasses with a low refractive index. In the visible spectrum, it is often quantified by the Abbe number that is defined by the relation,
(1)
where nC, nD, and nF are the refractive indices of the material at the following visible wavelengths: 656.3, 589.3, and 486.1 nm. While this factor is useful in the visible, for applications that require a broader spectral coverage, the complete dispersion curve n = f(λ) is more accurate and provides full information. Figure 3 shows these curves for fluorozirconate (ZFG) and fluoroindate (IFG) glasses. These data are useful for the design of optical fibers as their numerical aperture (NA) is typically calculated in the visible range of the spectrum, yet it changes at longer wavelengths.
FIG. 3.

Dispersion curves of standard IFG and ZFG fluoride glasses.

FIG. 3.

Dispersion curves of standard IFG and ZFG fluoride glasses.

Close modal
Material dispersion MD is defined as the second derivative of the dispersion curve with respect to wavelength,
(2)
MD has been reported for several fluoride glasses.27–31 The zero-dispersion wavelength (i.e., the wavelength where MD = 0) is an important factor that must be taken into account for the design of single mode fibers. Typical values are 1.5 μm for silica, 1.72 μm for ZBLAN, and 1.99 μm for IFG.27 

The refractive index also changes with temperature. While the numerical quantity dn/dT is positive for silica glass, it is negative for HMFGs. The dn/dT change in fluoride glasses has been addressed in the early stages of their development, and values range between −8 and −15 × 10−6 K−1.32 The main and common features of HMFGs are that dn/dT is negative and that the magnitude of the change is smaller as compared to other glasses such as chalcogenides.

The nonlinear refractive index is small in fluoride glasses, similar to fused silica. According to the literature, it is in the range 1−2 × 10−20 m2/W,23 which is substantially lower as compared to heavy metal oxide and chalcogenide glasses.

Numerous studies have investigated the chemical durability of fluorozirconate glasses.33,34 While HMFGs are resistant to fluorinating reagents, they are sensitive to water attack. Of main concern is their behavior in aqueous solutions. When a piece of glass is immersed in water, it undergoes a slow chemical attack. As a result, the surface is damaged and the pH of the solution drops, i.e., it becomes more acidic. The rate of attack increases rapidly as pH decreases, potentially resulting in a catastrophic runaway process. In contrast, the rate of attack can be kept very low in a slightly alkaline solution with a pH value of 8. Bulk glass is, therefore, stable in a buffered solution at this pH.35 The mechanism of attack is complex and includes several steps: first, fluorine anions are replaced by hydroxyls; then some glass components enter the liquid phase; finally, less soluble fluorides (e.g., BaF2) precipitate onto the glass surface, creating mechanical stress and surface defects.34 As a consequence, when liquid water stays at the fiber surface for several hours (even when the fiber is coated with an epoxy acrylate resin), the fiber becomes practically unusable.

In practice, one must absolutely avoid any liquid water remaining on the surface of an HMFG. It is important to note, however, that the rate of attack is relatively slow, which allows the use of water for polishing fluoride bulk glasses and preforms.

The durability of a glass depends on its chemical composition, in particular on the relative concentration of fluorides that exhibit significant solubility in water. For example, HMFGs containing sodium are more sensitive to water attacks, while calcium, yttrium, or thorium increase their resistance. When exposed to aqueous solutions, fluoroindates behave very similar to fluorozirconates, while fluoroaluminate glasses are more resistant in some compositions.21,23

There is no chemical reaction between HMFGs and water vapor below 100 °C. Bulk HMFGs and fibers do not suffer surface degradation when exposed to a humid atmosphere; optical components have been stored for tens of years in an ambient atmosphere without visible damage. However, condensation must be prevented. Deliquescence has never been observed for any current fluoride glasses.

HMFGs are classified as soft glasses in contrast to silica and soda lime glasses. Their microhardness depends on the chemical composition. The reported values are around 250 kg/mm2 for ZBLAN and fluoroindate glasses,32 yet can exceed 320 kg/mm2 for fluoroaluminates. Ultrasonic measurements are used to determine the three elastic moduli: bulk modulus, shear modulus, and Young’s modulus. A summary of the mechanical properties of ZFG, IFG, and AFG glasses in comparison to fused silica is provided in Table IV.

TABLE IV.

Mechanical moduli of optical fiberglasses. AFG: fluoroaluminate, ZFG: fluorozirconate, and IFG: indium fluoride glass. E is the Young’s modulus, G is the shear modulus, K is the bulk modulus, ν is the poison ratio, and Hk is the Knoop hardness.

E (GPa)G (GPa)K (GPa)ν []Hk (kg mm−2)
AFG 65 25 57 0.31 300–330 
ZFG 52–55 19–21 40–46 0.29–0.32 190–250 
IFG 44–55 19–21 47–52 0.31–0.33 180–230 
Silica 73 3100 36.70 0.17 770 
E (GPa)G (GPa)K (GPa)ν []Hk (kg mm−2)
AFG 65 25 57 0.31 300–330 
ZFG 52–55 19–21 40–46 0.29–0.32 190–250 
IFG 44–55 19–21 47–52 0.31–0.33 180–230 
Silica 73 3100 36.70 0.17 770 

Theoretical analysis suggests that the intrinsic strength of fluoride glasses should be half that of silica, and fracture toughness has been reported to be in the range from 0.25 to 0.30 MPa m1/2. In practice, mechanical failure usually originates from extrinsic defects, mostly surface defects. Because the emphasis is mainly on optical fibers, most available data relates to fibers. Depending on drawing conditions, fiber strengths may reach 1600 MPa,36,37 while typical values rather lie between 500 and 1000 MPa. This is still far below the intrinsic strength value.

The structure of HMFGs has been the subject of numerous studies. In analogy with other glasses, it is assumed that it can be described as an aperiodic network in which modifying cations are inserted. This network is constructed from large coordination polyhedra that are sharing corners and edges. Typical polyhedra are ZrF8 square antiprisms and dodecahedra, as well as AlF6 and InF6 octahedra. The structure of HMFGs does not comply with Zachariasen’s rules that specify that coordination polyhedra should be small—a tetrahedron or triangle—and should share only apexes, not edges.38 Alternatively, HMFG may be described as a disordered packing of fluorine anions in which cations are randomly inserted.39 Both models are equivalent, and the correspondence between them is illustrated in Fig. 4, which shows a bidimensional picture of a fluorozirconate glass.40 An additional element of the glass structure is the occurrence of vacancies in the random packing,41 as shown in Fig. 5. It leads to a “cluster” model and provides a physical meaning of the classical “excess free volume” concept that has been introduced a long time ago.

FIG. 4.

Bidimensional schematic of a fluorozirconate glass according to the random packing and the network models. Reproduced with permission from M. Poulain, “Glass systems and structures,” in Fluoride Glasses, Critical Reports on Applied Chemistry, edited by A. E. Comyns (John Wiley & Sons, Chichester, New York, 1989), pp. 11–48. Copyright 1989 Wiley.

FIG. 4.

Bidimensional schematic of a fluorozirconate glass according to the random packing and the network models. Reproduced with permission from M. Poulain, “Glass systems and structures,” in Fluoride Glasses, Critical Reports on Applied Chemistry, edited by A. E. Comyns (John Wiley & Sons, Chichester, New York, 1989), pp. 11–48. Copyright 1989 Wiley.

Close modal
FIG. 5.

2D representation of the structure of ZBLAN glass according to the lacunar model of the vitreous state. Reproduced with permission from M. Poulain, “Vacancy model of glass,” in Proceedings 10th ISNOG Symposium (Corning, New York, 1996), Vol. B, pp. 10–15. Copyright 2005 Wiley.

FIG. 5.

2D representation of the structure of ZBLAN glass according to the lacunar model of the vitreous state. Reproduced with permission from M. Poulain, “Vacancy model of glass,” in Proceedings 10th ISNOG Symposium (Corning, New York, 1996), Vol. B, pp. 10–15. Copyright 2005 Wiley.

Close modal

This picture emphasizes the existence of bidimensional channels that account for ionic conductivity. Despite these vacancies, the structure of HMFGs is highly compact, in stark contrast to silica or germania glasses, in which the vacancy rate is close to 50% as it is impossible to fill the entire space only with tetrahedra sharing apexes.

HMFGs belong to the group of low melting point glasses. Their characteristic and processing temperatures lie below 1000 °C. The glass transition temperature Tg marks the limit between the solid and the molten (liquid) state. In practice, glass components cannot be used beyond this temperature. Most HMFGs are prone to devitrification when heated between the glass transition and melting temperatures. The crystallization temperature Tx is usually measured by thermal analysis (Differential Thermal Analysis, DTA, or Differential Scanning Calorimetry, DSC) at a heating rate of 10 K/min. The thermal stability range is defined as ΔT = Tx − Tg and gives an estimate of glass stability vs devitrification. Table V lists the characteristic temperatures of various HMFGs.

TABLE V.

Physical properties of HMFGs with compositions as per Table I. Tg and Tx are the temperatures for glass transition and that of crystallization, respectively. ρ is the density, n is the refractive index, and α is the coefficient of thermal expansion.

GlassTg (°C)Tx (°C)ρ (g/cm3)nα (10−7 K−1)
ZBLA 320 392 4.54 1.519 168 
HBLA 325 426 5.88 1.504 165 
ZBNA 255 340 4.28 1.497 200 
ZBLAN 263 370 4.52 1.498 200 
ZBLALi 254 369  1.516 202 
ZBLANP 251 369  1.517 200 
ZBLALiP 245 352  1.5295 202 
IYB 333 416 4.96   
IZBS 301 388 5.14 1.5005 187 
IZBSC 292 383 5.10 1.498 182 
IZSBN 291 380 4.96 1.487 192 
PIGCZ 248 360 6.23 1.602 165 
YABC 430 535 4.00 1.440 165 
BATY 446 535 5.10 1.487 147 
AFG3H 392 490 3.85 1.439 152 
ABZYZn 373 506 4.49 1.488 150 
AZBZnYC 340 467 4.32 1.486 164 
GlassTg (°C)Tx (°C)ρ (g/cm3)nα (10−7 K−1)
ZBLA 320 392 4.54 1.519 168 
HBLA 325 426 5.88 1.504 165 
ZBNA 255 340 4.28 1.497 200 
ZBLAN 263 370 4.52 1.498 200 
ZBLALi 254 369  1.516 202 
ZBLANP 251 369  1.517 200 
ZBLALiP 245 352  1.5295 202 
IYB 333 416 4.96   
IZBS 301 388 5.14 1.5005 187 
IZBSC 292 383 5.10 1.498 182 
IZSBN 291 380 4.96 1.487 192 
PIGCZ 248 360 6.23 1.602 165 
YABC 430 535 4.00 1.440 165 
BATY 446 535 5.10 1.487 147 
AFG3H 392 490 3.85 1.439 152 
ABZYZn 373 506 4.49 1.488 150 
AZBZnYC 340 467 4.32 1.486 164 

The viscosity η is an important physical parameter in glass technology. It is measured between the glass transition temperature and the liquidus temperature. At Tg, it is close to 1013 poises, while it may be lower than 1 poise at the liquidus temperature. The variation, therefore, covers more than ten orders of magnitude. For this reason, it is typically plotted on a logarithmic scale, i.e., Log η = f (T−1). This plot is linear in the case of an Arrhenian behavior that is exceptional for glasses and applies only to pure silica and vitreous BeF2. While most industrial glasses follow the Vogel–Tammann–Fulcher law, the viscosity of fluoride glasses deviates even further from linearity.42,43 They, therefore, belong to the group of “fragile glasses” as introduced by Angell.44 A typical viscosity/temperature curve for fluorozirconate glasses with different sodium concentrations42 is shown in Fig. 6. The viscosity decreases sharply above the glass transition temperature and is very low around the liquidus temperature (less than 1 poise). As a consequence, the temperature must be controlled very accurately for glass transformation, which is of utmost importance for fiber drawing.

FIG. 6.

Viscosity as a function of temperature for fluorozirconate glasses with different sodium concentrations (0, 10, and 20 mol % NaF). Glass compositions: ZBLA from Table I; ZBLAN20 is ZBLAN in Table I. ZBLAN10: 55 ZrF4, 28 BaF2, 4LaF3, 3 AlF3, and 10 NaF. Adapted from Ref. 42.

FIG. 6.

Viscosity as a function of temperature for fluorozirconate glasses with different sodium concentrations (0, 10, and 20 mol % NaF). Glass compositions: ZBLA from Table I; ZBLAN20 is ZBLAN in Table I. ZBLAN10: 55 ZrF4, 28 BaF2, 4LaF3, 3 AlF3, and 10 NaF. Adapted from Ref. 42.

Close modal
When heated beyond the glass transition temperature, most HMFGs generate crystals in the molten state, and their devitrification has been extensively studied. It involves two different mechanisms: nucleation and crystal growth. They are investigated using thermal analysis at temperatures higher than Tg to quantify the relative amount of the crystalline phase. They are based on the Johnson–Melh–Avrami–Kolmogorov (JMAK) relation,45–47 
(3)
where x is the crystalline fraction, t is the time, k is a constant that depends on nucleation and crystal growth, and n is the Avrami exponent. This exponent can take different values depending on the nucleation and growth mechanisms. This relation applies to the isothermal stage. It is assumed that the constant k changes with temperature T according to
(4)

Three kinetical parameters are derived from the JMAK relation: n, E, and k0. The physical meaning of the activation energy E and the frequency constant k0 is not clear. In multicomponent glasses, the chemical composition of the crystalline phases that are the result of devitrification is generally different from that of the parent glass. The search for stable glass compositions remains empirical, as there is no obvious way to predict glass stability. The so-called “confusion principle” has often been used as a guideline to find glass compositions that are more stable against devitrification. This principle states that glass stability should increase with the number of glass components: a quaternary glass should be more stable than a ternary glass, which is expected to be more stable than a binary one. The standard ZBLAN fluorozirconate glass has been defined this way.

The potential of HMFG fibers for telecommunications has already been pointed out in their early stages of development when it was emphasized that the intrinsic optical losses of these fibers could reach values that are well below those of silica fibers, i.e., as low as 10−2–10−3 dB/km. This would enable repeater-less long haul links, possibly even transoceanic,18,48–50 a potential that raised enormous interest among major telecom companies. However, after a few years of intense research, it appeared that the optical losses could not be reduced below the levels achieved in silica fibers, where the theoretical loss limits had been reached. The major problems were identified to be chemical impurities and devitrification. More recently, it was reported that crystallization could be inhibited in microgravity. Several US companies currently investigate this approach.51 

Optical amplification is another field of application of HMFG fibers. This applies specifically to the telecommunications O and S bands for which rare earth doped silica fibers do not allow amplification. Fluoride glasses that exhibit low phonon energy make it possible to develop such optical amplifiers using thulium and praseodymium as active elements.

Combining individual telescopes via fiber links significantly increases the achievable accuracy and resolution. The atmospheric transmission window between 2–2.5 μm (known as the astronomical K band) is highly attractive in this respect. The combination of two signals from distinct telescopes results in a broadband interferometric pattern from which images are obtained. This approach requires low loss single mode fibers that can be manufactured only from fluoride glasses. After pioneering achievements at the Kitt Peak Observatory in the US,52 the method was used for several instruments: Ohana,53 GRAVITY,54,55 SPIRou,56 and SPIP.57 Gravity was used by the 2021 Nobel laureates in physics.

Fiber links are required for many systems operating in the mid-IR, for example, for sensing and remote detection. Fluoride glass fibers are used for wavelengths between 2.4 μm (the limit of silica fibers) and 5 μm. The various fields of application include scientific research, thermal imaging, and laser power delivery, most importantly around 2.9 μm, which coincides with the absorption peak of liquid water. With the emergence of Quantum Cascade Lasers (QCLs) and Interband Cascade Lasers (ICLs) fluoride glass optical fibers that can be coupled to these diodes are in high demand.

Another field of applications is enabled by the excellent UV transparency of fluoride fibers. Compared to silica fibers, HMFG fibers are less sensitive to photodarkening, and this makes them well suited for the delivery of UV light.

Rare earth ions have long been studied for laser emission, and various solid-state laser crystals have been used for this purpose. In comparison to crystals, actively doped optical fibers offer a range of advantages, especially long interaction lengths and improved heat management. Among other parameters, the possibility of laser emission from a particular atomic transition depends on the phonon energy of the host material. While it is rather low in most crystals, it is higher in classical oxide glasses, which reduces the number of available emission lines. Rare earth doped fluoride fibers, on the other hand, combine low phonon energy, low optical losses, and a high rare earth solubility. As a consequence, a large number of new emission lines have been identified. This concerns not only the mid-infrared beyond 2 μm58 but also visible lasers either by upconversion or by pumping using blue (GaN) diodes. Continuous-wave (cw) fluoride glass fiber lasers operating at 2.9 μm are now being industrially produced for use in dermatology, where they are replacing more traditional CO2 lasers as they are more efficient at low power and less painful for the patient.59 Those lasers are also very promising for endoscopic surgery.

Multi-watt visible fiber lasers are currently under development. Based on rare earth-doped fluoride fibers, they will find important applications in several medical fields such as ophthalmology, microscopy, imaging, and diagnosis.

Based on nonlinear optical effects, fibers can be used to generate broadband emission that is fundamentally only limited by the transparency window of the fiber material. Such supercontinuum (SC) sources typically consist of a pulsed laser source with high peak power that is injected into a low loss optical fiber. Using silica fibers, SC sources can cover a wide spectrum between 600 and 2500 nm. Based on fluoride fibers, SC sources with a wavelength range between 700 and 5000 nm are available.60 These sources are mainly developed for mid-IR spectroscopy but can also be used for infrared countermeasures.

The concentration of fluorescent rare earth ions in glasses can be adjusted with high precision. The low phonon energy of fluoride glasses further enhances their efficiency, which makes them suitable for fluorescence standards in the visible spectrum (green, yellow, and red).

Fluoride glasses have been extensively studied and developed over the last fifty years. Initially, the prospect of ultra-low optical losses that could enable repeater-less long-haul transmission attracted significant interest and funding. Later, fluoride optical fibers were found to be complementary to silica fibers for mid-infrared transmission and fiber lasers.

A large number of multicomponent fluoride glasses have been reported in the past, yet only a small fraction has been further developed for practical applications. Those can be categorized into the three major groups of fluoroaluminates, fluorozirconates, and fluoroindates which are recognized as an original class of glasses with unusual structure, low characteristic temperatures, and specific physical properties. The main interest in these glasses stems from their broad optical transmission that extends from the UV to the mid-infrared spectrum and their extremely low OH contamination.

Most applications of HMFGs are based on optical fibers. Fiber drawing requires glass compositions that are very stable against devitrification. To date, a large variety of fibers have been reported, including multimode, singlemode, polarization maintenance, and double clad. Passive applications for fluoride fibers include sensing, interferometry, and laser power delivery, whereas active applications include rare earth-doped fibers, leading to new fiber lasers in the mid-IR and visible spectral ranges.

In order to be used as materials for optical fibers, HMFGs must withstand the processing conditions and be resistant enough to devitrification. Their extended optical transmission range makes them suitable for passive broadband optical components. Their low phonon energy allows the development of mid-infrared fiber lasers and supercontinuum sources. Numerous sensors and analysis systems based on fluoride fibers will emerge in the coming years. One can expect important applications in the medical field.

The author has no conflicts to disclose.

Marcel Poulain: Writing – original draft (equal).

Data sharing is not applicable to this article, as no new data were created or analyzed in this study.

1.
M.
Poulain
,
M.
Poulain
, and
J.
Lucas
, “
Verres fluores au tetrafluorure de zirconium proprietes optiques d’un verre dope au Nd3+
,”
Mater. Res. Bull.
10
(
4
),
243
246
(
1975
).
2.
A. E.
Comyns
, “
Fluoride glasses
,”
Critical Reports on Applied Chemistry
(
John Wiley & Sons
,
Chichester
,
1989
), Vol.
27
.
3.
I. D.
Aggarwal
and
G.
Lu
,
Fluoride Glass Fiber Optics
(
Academic Press
,
1991
).
4.
M.
Poulain
,
S.
Cozic
, and
J. L.
Adam
, “
Chapter 2 - Fluoride glass and optical fiber fabrication
,” in
Mid-Infrared Fiber Photonics
,
Woodhead Publishing Series in Electronic and Optical Materials
, edited by
S.
Jackson
,
M.
Bernier
, and
R.
Vallée
(
Woodhead Publishing
,
2022
), pp.
47
109
.
5.
C. M.
Baldwin
,
R. M.
Almeida
, and
J. D.
Mackenzie
, “
Halide glasses
,”
J. Non-Cryst. Solids
43
(
3
),
309
344
(
1981
).
6.
P. W.
France
,
M. G.
Drexhage
,
J. M.
Parker
,
M. W.
Moore
,
S. F.
Carter
, and
J. V.
Wright
,
Fluoride Glass Optical Fibres
(
Blackie
,
Glasgow and London: P. W. France
,
1990
).
7.
S.
Jackson
et al, “
Chapter 7 - Spectroscopy of the rare-earth-ion transitions in fluoride glasses
,” in
Mid-Infrared Fiber Photonics
,
Woodhead Publishing Series in Electronic and Optical Materials
, edited by
S.
Jackson
,
M.
Bernier
, and
R.
Vallée
(
Woodhead Publishing
,
2022
), pp.
333
399
.
8.
S.
Takahashi
,
T.
Shibata
,
T.
Kanamori
, and
S.
Mitachi
, “
New fluoride glasses for IR transmission
,” in
Physics in Fiber Optics
,
Advances in Ceramics
(
American Ceramic Society
,
1981
), Vol.
2
, pp.
74
82
.
9.
M. J.
Poulain
,
M. A.
Poulain
, and
G.
Maze
, “
Nouveaux verres fluorés
,” EP0036373A1, Sep. 23, 1981 (Online). Available: https://patents.google.com/patent/EP0036373A1/fr?inventor=Marcel+Poulain&sort=new&page=1 (accessed 18 Nov 2019).
10.
J.-P.
Miranday
,
C.
Jacoboni
, and
P. R.
De
, “
Glasses containing fluorine, their preparation and their application
,” US4328318 (A), May 04, 1982 (Online). Available: https://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=19820504&DB=&locale=en_EP&CC=US&NR=4328318A&KC=A&ND=4 (accessed 16 October 2019).
11.
J. J.
Videau
,
B.
Dubois
, and
J.
Portier
, “
Verres à base de fluorure d’indium
,”
C. R. Hebd. Seances Acad. Sci.
297
(
II
),
483
(
1983
).
12.
M. J.
Poulain
,
M. A.
Poulain
,
Y.
Messaddeq
, and
A.
Soufiane
, “
Fluoroindate glasses
,” in
Solid-State Optical Materials
,
Ceramic Transactions
, edited by
A.
Bruce
and
B.
Hiremath
(
American Ceramic Society
,
Westerville, OH
,
1992
), Vol.
28
, pp.
381
392
.
13.
P. P.
Fedorov
,
R. M.
Zakalyukin
,
L. N.
Ignat’eva
, and
V. M.
Bouznik
, “
Fluoroindate glasses
,”
Russ. Chem. Rev.
69
(
8
),
705
716
(
2000
).
14.
J.-P.
Miranday
,
C.
Jacoboni
, and
R.
De Pape
, “
New transition metal fluoride glasses isolated in the PbF2 MtIIF2 MtIIIF3 systems
,”
J. Non-Cryst. Solids
43
,
393
401
(
1981
).
15.
M.
Poulain
, “
Halide glasses
,”
J. Non-Cryst. Solids
56
(
1
),
1
14
(
1983
).
16.
Y.
Messaddeq
and
M.
Poulain
, “
Stabilizing effect of aluminium, yttrium and zirconium in divalent fluoride glasses
,”
J. Non-Cryst. Solids
140
,
41
46
(
1992
).
17.
A.
Lecoq
and
M.
Poulain
, “
Etude du rôle stabilisateur de l’aluminium dans les verres au tétrafluorure de zirconium
,”
Verres Réfract.
34
(
3
),
333
342
(
1980
).
18.
M. G.
Drexhage
,
B.
Bendow
, and
C. T.
Moynihan
, “
IR transmitting fluoride glasses
,”
Laser Focus
10
,
62
65
(
1980
).
19.
K.
Ohsawa
and
T.
Shibata
, “
Preparation and characterization of ZrF4-BaF2-LaF3-NaF-AIF3 glass optical fibers
,”
J. Lightwave Technol.
2
(
5
),
602
606
(
1984
).
20.
T.
Kanamori
,
K.
Oikawa
,
S.
Shibata
, and
T.
Manabe
, “
BaF2-CaF2-YF3-AlF3 glass systems for infrared transmission
,”
Jpn. J. Appl. Phys.
20
(
5
),
L326
L328
(
1981
).
21.
M.
Poulain
,
M.
Poulain
, and
M.
Matecki
, “
Verres fluores a large bande de transmission optique et a haute resistance chimique
,”
Mater. Res. Bull.
16
(
5
),
555
564
(
1981
).
22.
M.
Poulain
,
M.
Poulain
, and
M.
Matecki
, “
Rare earths in fluoride glasses: The LnF3 AlF3 ThF4 BaF2 systems and related glasses
,”
J. Non-Cryst. Solids
51
(
2
),
201
215
(
1982
).
23.
T.
Izumitani
,
T.
Yamashita
,
M.
Tokida
,
K.
Miura
, and
H.
Tajima
, “
New fluoroaluminate glasses and their crystallization tendencies and physical-chemical properties
,”
Mater. Sci. Forum
19–20
,
19
26
(
1987
).
24.
M.
Poulain
, “
Properties and applications of fluoride glasses
,” in
Non Crystalline Materials for Optoelectronics
, edited by
M.
Popescu
(
INOE
,
Bucharest
,
2004
), pp.
335
361
.
25.
Y.
Messaddeq
,
A.
Delben
,
M.
Boscolo
,
M. A.
Aegerter
,
A.
Soufiane
, and
M.
Poulain
, “
New fluoroindate glass compositions
,”
J. Non-Cryst. Solids
161
,
210
212
(
1993
).
26.
M.
Poulain
and
G.
Mazé
, “
On the Chemistry of fluoride glasses
,”
Chemtronics
3
,
77
85
(
1988
).
27.
F.
Gan
, “
Optical properties of fluoride glasses: A review
,”
J. Non-Cryst. Solids
184
,
9
20
(
1995
).
28.
L.
Jeunhomme
,
H.
Poignant
, and
M.
Monerie
, “
Material dispersion evaluation in a fluoride glass
,”
Electron. Lett.
17
(
21
),
808
809
(
1981
).
29.
B.
Bendow
,
R. N.
Brown
,
M. G.
Drexhage
,
T. J.
Loretz
, and
R. L.
Kirk
, “
Material dispersion of fluorozirconate-type glasses
,”
Appl. Opt.
20
(
21
),
3688
3690
(
1981
).
30.
S.
Mitachi
and
T.
Miyashita
, “
Refractive-index dispersion for BaF2–GdF3–ZrF4–AlF3 glasses
,”
Appl. Opt.
22
(
16
),
2419
2425
(
1983
).
31.
A.
Soufiane
,
F.
Gan
,
H.
L’Helgoualch
, and
M.
Poulain
, “
Material dispersion in optimized fluoroindate glasses
,”
J. Non-Cryst. Solids
184
,
36
39
(
1995
).
32.
T.
Izumitani
,
T.
Yamashita
,
M.
Tokida
,
K.
Miura
, and
H.
Tajima
, “
Physical, chemical properties and crystallization tendency of the new fluoroaluminate glasses
,” in
Halide Glasses for Infrared Fiberoptics
,
NATO ASI Series
, edited by
R. M.
Almeida
(
Springer
,
Netherlands, Dordrecht
,
1987
), pp.
187
197
.
33.
C. J.
Simmons
and
J. H.
Simmons
, “
Chemical durability of fluoride glasses: I, reaction of fluorozirconate glasses with water
,”
J. Am. Ceram. Soc.
69
(
9
),
661
(
1986
).
34.
R. H.
Doremus
,
N. P.
Bansal
,
T.
Bradner
, and
D.
Murphy
, “
Zirconium fluoride glass: Surface crystals formed by reaction with water
,”
J. Mater. Sci. Lett.
3
(
6
),
484
488
(
1984
).
35.
C. J.
Simmons
, “
Chemical durability of fluoride glasses: III, the effect of solution pH
,”
J. Am. Ceram. Soc.
70
,
654
661
(
1987
).
36.
H. W.
Schneider
,
A.
Schoberth
,
A.
Staud
, and
C.
Gerndt
, “
Optical and mechanical properties of fluoride glass fibers for infrared transmission systems Schneider
,”
Siemens Forsch.- Entwicklungsber.
17
(
3
),
147
153
(
1988
).
37.
H. W.
Schneider
, “
Fiber drawing
,” in
Critical Reports on Applied Chemistry
, edited by
A. E.
Comyns
(
John Wiley & Sons
,
Chichester, New York
,
1989
), Vol. 27, pp.
185
199
.
38.
W. H.
Zachariasen
, “
The atomic arrangement in glass
,”
J. Am. Chem. Soc.
54
,
3840
(
1932
).
39.
M.
Poulain
, “
Glass formation in ionic systems
,”
Nature
293
(
5830
),
279
280
(
1981
).
40.
M.
Poulain
, “
Glass systems and structures
,” in
Fluoride Glasses
,
Critical Reports on Applied Chemistry
, edited by
A. E.
Comyns
(
John Wiley & Sons
,
Chichester, New York
,
1989
), Vol. 27, pp.
11
48
.
41.
M.
Poulain
, “
Vacancy model of glass
,” in
Proceedings 10th ISNOG
(
Corning
,
New York
,
1996
), Vol.
B
, pp.
10
15
.
42.
W. C.
Hasz
and
C. T.
Moynihan
, “
Physical properties of ZrF4-based glass-forming melts and glasses
,”
J. Non-Cryst. Solids
140
,
285
292
(
1992
), Also: Mater. Sci. Forum 32–33, 589–594.
43.
T.
Yano
,
S.
Inoue
,
H.
Kawazoe
, and
M.
Yamane
, “
Viscosity of the melt of ZrF4-BaF2-LaF3-AlF3-NaF system
,”
Mater. Sci. Forum
67–68
,
563
568
(
1991
).
44.
C. A.
Angell
, “
Relaxation in liquids, polymers and plastic crystals—Strong/fragile patterns and problems
,”
J. Non-Cryst. Solids
131–133
,
13
31
(
1991
).
45.
W. A.
Johnson
and
R. F.
Mehl
, “
Reaction kinetics in processes of nucleation and growth
,”
Trans. Am. Inst. Min. Metall. Pet. Eng.
135
,
416
442
(
1939
).
46.
M.
Avrami
, “
Kinetics of phase change: I general theory
,”
J. Chem. Phys.
7
(
12
),
1103
1112
(
1939
).
47.
A. N.
Kolmogorov
, “
On statistical theory of crystallization of metals
,”
Bull. Acad. Sci. URSS (Sci. Mat. Nat.)
3
,
355
359
(
1939
).
48.
J. R.
Gannon
, “
Optical fibre materials for operating wavelengths larger than 2 µm
,”
J. Non-Cryst. Solids
42
,
239
246
(
1980
).
49.
S.
Shibata
,
M.
Horiguchi
,
K.
Jinguji
,
S.
Mitachi
,
T.
Kanamori
, and
T.
Manabe
, “
Prediction of loss minima in infra-red optical fibres
,”
Electron. Lett.
17
(
21
),
775
777
(
1981
).
50.
S. F.
Carter
,
J. R.
Williams
,
M. W.
Moore
,
D.
Szebesta
, and
S. T.
Davey
, “
Prospects for ultra-low-loss fluoride fibres at BTRL
,”
J. Non-Cryst. Solids
140
,
153
158
(
1992
).
51.
I.
Cozmuta
,
S.
Cozic
,
M.
Poulain
,
S.
Poulain
, and
J. R. L.
Martini
, “
Breaking the silica ceiling: ZBLAN-based opportunities for photonics applications
,” in
Optical Components and Materials XVII
(
International Society for Optics and Photonics
,
2020
), p.
112760R
.
52.
V.
Coudé du Foresto
and
S. T.
Ridgway
, “
Fluor - A Stellar interferometer using single-mode fibers
,” in High Resolution Imaging by Interferometry II, edited by
J.
Beckers
and
F.
Merkle
(ESO, Garching,
1991
), pp.
731
740
.
53.
T.
Kotani
,
G.
Perrin
,
S.
Vergnole
,
J.
Woillez
, and
J.
Guerin
, “
Characterization of fluoride fibers for the optical Hawaiian array for nanoradian astronomy project
,”
Appl. Opt.
44
(
24
),
5029
5035
(
2005
).
54.
K.
Perraut
et al, “
Single-mode waveguides for GRAVITY: I. The cryogenic 4-telescope integrated optics beam combiner
,”
A&A
614
,
A70
(
2018
).
55.
G.
Perrin
et al, “
Single-mode waveguides for GRAVITY: II. Single-Mode fibers and fiber control unit
,”
A&A
681
,
A26
(
2024
).
56.
Y.
Micheau
et al, “
SPIRou @ CFHT: Fiber links and pupil slicer
,”
SPIE Proc.
8446
,
84462R
(
2012
).
57.
S.
Baratchart
et al, “
SPIP at TBL, the faithfull companion of the SPIRou spectropolarimeter at CFHT: Integration, tests and performances
,”
Proceedings SPIE
(SPIE, 2022),
12184
.
58.
S. D.
Jackson
and
R. K.
Jain
, “
Fiber-based sources of coherent MIR radiation: Key advances and future prospects (invited)
,”
Opt. Express
28
(
21
),
30964
31019
(
2020
).
59.
T. N.
Murray
,
J. K.
Hu
, and
P. M.
Friedman
, “
Full-face and neck resurfacing with a novel ablative fractional 2910 nm erbium-doped fluoride glass fiber laser for advanced photoaging
,”
Lasers Surg. Med.
56
(
3
),
249
256
(
2024
).
60.
J.-C.
Gauthier
et al, “
Mid-IR supercontinuum up to 5.4 μm in a 15 meter-long fluoroindate fiber
,” in
Frontiers in Optics
(
Optical Society of America
,
2016
), paper FTu1I.2.