Borate glasses for scientific and industrial applications: a review

Abstract

Research in borate glasses has started as a scientific curiosity and as an aid to explain the structure of oxide glasses in general. This effort led to a better understanding of the structure and unique properties of borate glasses. Although silicate and borosilicate glasses satisfy the vast majority of scientific and industrial needs, there are certain circumstances where they are not satisfactory. Furthermore, borate glasses offer certain advantages over silicate glasses which are not well known, neither well explored. However, certain characteristics of borate glasses such as their affinity to water requires that they are well selected, designed, or developed to satisfy the specifications of a given application. This review aims to explore and report some of the key properties of binary borate glasses. It also provides certain guidelines for several applications where the scientific literature is discussed in connection with industrial experience and technological needs.

Introduction

The continuous effort for scientific and technological development is a major driving force for the search of materials with improved, tailored, or optimized properties. Although silica glasses satisfy the majority of requirements for a wide range of applications, borate glasses are also getting increased attention because of their unique properties. An important advantage of borate glasses over silica glasses is their significantly lower melting and softening temperatures. Pure B₂O₃ has a glass transition temperature (T _g) of ~260 °C [1] and a melting temperature (T _m) of ~450 °C [2] which are much lower than the corresponding temperatures of T _g ≅ 1100 °C and T _m ≅ 1728 °C for SiO₂ [1, 2]. Borate glass is normally not useful for any application in the pure form because of its very low chemical durability and high affinity for water. Therefore, it is used in combination with other oxides such as Al₂O₃ or SiO₂ which leads to improved chemical durability yet at the cost of higher processing temperatures. Nevertheless, certain borate glasses of practical use still have considerably lower melting temperatures compared to silica-based glasses. The low processing temperatures are important in applications like sealing and passivation of electronic instruments [3–5].

The addition of B₂O₃ to SiO₂ glasses led to the development of borosilicate glasses with a broad range of industrial applications. These glasses have higher chemical durability thermal shock resistance, and electrical resistivities compared to soda-lime glasses. The best known example is Pyrex^® glass which is used for labware and consumer products such as glass plates and oven ware. The borate content in commercial borosilicate glasses ranges between 13 and 25 wt% [1].

Borate glasses are preferred over silica glasses in sealing applications where the reactivity of silica poses a threat for certain components of the system. For example, silica reacts with sodium in sodium vapor lamps and with lithium available in liquid electrolytes of lithium batteries. Recent studies that aim to develop glass sealants for solid oxide fuel cells (SOFCs) also prefer Si-free glass chemistry since silicon diffusion from silica glass into yttria stabilized zirconia (YSZ) electrolytes has the potential to poison the YSZ and reduce its ionic conductivity [6].

It is never sufficient to consider a single parameter for a particular application. For example, sealing applications for electronics require low softening temperatures of the sealing glass in order to minimize damage to electronic components and to reduce production costs. In addition, a glass material used for sealing applications must exhibit a good thermal expansion match and limited chemical reactivity with the base material, strong interfacial bonding, and mechanical strength. Furthermore, certain applications such as SOFC’s, aerospace, and biomedicine require hermetic (gas and liquid tight) sealing [7]. The interdependence of certain properties complicates the material selection or product development process. As commonly happens in most design and development activities, the improvement of one parameter could be detrimental to another one. For example, in sealing applications, reducing the melting temperature of the sealing glass may result in reduced chemical durability.

The aim of this review is to bring together available data from the literature, to review some structure–property relationships and to make some extrapolations, interpretations, and recommendations for the optimization and selection of borate glass compositions for selected applications.

Scientific and industrial applications of borate glasses

This section will specifically consider applications of non-silicate glasses based on borates. Critical physical and chemical aspects related to some of these applications will be discussed. Although borosilicate glasses have many important applications, they will not be considered here specifically because the focus of this paper is glass compositions which offer unique properties due to the borate network or borate chemistry.

Sodium vapor lamps, sodium beta batteries, and lithium batteries

Several applications of borate glasses were developed because silica in silicate glasses can be harmed by the chemistry of the application in question. Sodium vapor lamps, sodium-beta batteries, and Li/SO₂ batteries are three such applications. The common denominator of these applications is the alkali element which attacks silica in silicate glasses.

Low-pressure sodium vapor lamps were developed by Phillips and first applied in England to illuminate a road in south London near Phillips Company’s research labs [8]. Hot sodium vapor is highly reactive. It attacks and blackens silica-containing glasses. This problem was solved with the development of aluminoborate glasses. However, aluminoborate glasses are attacked by atmospheric moisture, are more expensive, and more difficult to shape. Therefore, only a thin aluminoborate glass coating is currently used inside ordinary soda-lime glass tubing [8, 9]. High-pressure sodium lamps were introduced in the 1970s which employed translucent alumina arc tubes. These lamps had certain advantages over low-pressure lamps such as smaller size and lower cost. Nevertheless, low-pressure sodium lamps are more efficient and they are still being used for street lighting [9].

Sodium beta batteries are rechargeable high-temperature batteries which employ liquid sodium as the active metal in the negative electrode and β″-Al₂O₃ as the electrolyte. One type of sodium beta battery is the sodium/sulfur battery which was introduced in the 1970s. Although this battery was considered for electric-vehicle applications, development programs have ceased in the mid-1990s. However, the sodium/sulfur technology has been utilized by Japanese developers for stationary energy storage applications [10]. In a typical design for a tubular cell battery, a metal case contains the liquid sodium electrode. The β″-Al₂O₃ canister acts as the ion-conducting solid electrolyte which also contains the liquid sulfur electrode [11]. An α-Al₂O₃ ring surrounding the β″-Al₂O₃ tube is bonded by a layer of sealing glass. This glass must resist the corrosive effect of both liquid sodium and sulfur electrodes at the operation temperatures of 290–390 °C.

Another group of sodium beta batteries are sodium/metal halide batteries. Sodium/metal halide technologies were introduced a decade after sodium/sulfur batteries. These technologies offer several advantages over sodium/sulfur battery technologies. Instead of molten sulfur, a metal halide such as nickel chloride or iron chloride is used in the solid state.

Requirements for the sealing glass in sodium beta batteries include [12]

• corrosion resistance against liquid sodium and metal halides,

• hermetic sealing,

• thermal expansion match with α- and β″-Al₂O₃, and

• high bond strength of glass to ceramic joints.

During the initial stages of development, it was realized that silica glasses are unstable against liquid sodium. This problem was addressed by extensive research programs to develop alternative sealing glasses [11]. Most of the work by different groups led to the development of aluminoborate glasses for these applications. Although the amount of glass material used in battery applications is relatively small, sealing glasses are one of the key enablers of high-performance batteries. Some of the important sealing glasses used in battery and other applications are summarized in Table 1.

Table 1 Commercial glass types used for sealing applications

There are various types of lithium batteries produced for the consumer market and for industrial applications. The most common type used in the consumer market uses lithium metal as the anode and manganese dioxide as the cathode. This type of lithium battery does not require a glass to metal hermetic seal because it does not involve any pressure of reactant materials. Some types of lithium batteries involve soluble cathodes such as sulfur dioxide (SO₂) or thionyl chloride (SOCl₂) and nonaqueous liquid electrolytes such as lithium bromide/acetonitrile (in the case of an SO₂ cathode) and lithium tetrachloroaluminate (in the case of an SOCl₂ cathode). The reaction of the cathode and electrolyte materials often causes pressure buildup in the battery cell. Hermetic glass-to-metal seals are commonly used to contain the pressure rise in the cell and to isolate the anode from the cathode [10, 13–15]. Silicate glasses are not a good choice for this application because silica reacts with lithium and forms conductive silicides. Such reactions may both weaken the seal mechanically and cause an electrical short circuit between the battery case (anode) and the pin (cathode). TA-23 glass used for Li ambient temperature batteries contains about 45 wt% silica (Table 1) and resists corrosion for 5–10 years. Alkaline earth aluminoborate glasses have been developed for such applications since they resist attack by alkali metals. CABAL-12 has better corrosion resistance since it does not contain any silica. Glasses used for these applications should have the following properties:

• a good thermal expansion match to the metal components involved,

• good adhesion to the metal components involved,

• corrosion resistance to lithium,

• electrical insulation, and

• a low sealing temperature.

Figure 1 shows a common seal design used in lithium batteries and other similar electrochemical cells. The electrical lead pin is usually made from materials such as molybdenum, tungsten, or tantalum. In addition to being refractory, a common feature of these metals is their low thermal expansion coefficient. CTE values of pure Mo, W, and Ta are listed as 5 × 10⁻⁶/°C, 5 × 10⁻⁶/°C, and 7 × 10⁻⁶/°C, respectively [16]. This level of thermal expansion is easier to match by commercial sealing glasses like TA-23 and CABAL-12, which have CTE values in the order of 6 × 10⁻⁶/°C (Table 1). However, alternative borate glasses with higher CTE values have also been developed to match the higher thermal expansion of certain pin materials such as 446 stainless steel and Alloy-52 with CTE values of 11.4 × 10⁻⁶/°C and 9.8 × 10⁻⁶/°C, respectively. These are boroaluminate glasses with SrO and BaO substituted for the CaO and MgO used in CABAL-12 [17].

Fig. 1

Cross section of a common seal design used for lithium batteries and other similar electrochemical cells

Sealing of the viscous glass to the metal occurs at high temperatures leading to the formation of an oxide layer on metal components. Usually this oxide layer is an essential part of a hermetic seal. Glasses with B₂O₃ content above 15 % are known to produce strong adhesion to metals due to the low surface tension of B₂O₃ and a suitable oxide layer formed on the metal. The persistence of adhesion also depends on the thermal expansion match. Even if there is no ideal match between the CTEs of the glass and metal, various design strategies can eliminate the possibility of stress buildup or destruction of the interface bond [18].

The seal must be an electrical insulator to prevent electric flow between the terminal pin and the header body. This requirement is readily satisfied by borate sealing glasses.

Solder glasses

Another major field of application is low temperature sealing glasses also called solder glasses or frit glasses. These glasses have a softening point below 550 °C. They are employed for sealing, packaging, passivation, and enameling. Industrial sectors which require solder glasses are mainly electronics, electrooptics, mechatronics, and electrical industries. Frit glass is used in the form of glass powder with some polymeric binder and can be applied as a paste. It is also an important material for jewelry, decoration, and ceramics industries. In electronics packaging industries, solder glasses are used for sealing vacuum fluorescent displays, lamp bulbs, plasma display panels, aluminum IC packages, and ferrites. They are also used as insulation coating and overcoating, and for fixing optical communication devices. Almost all commercial solder glasses used to contain lead until recently (Table 1) but the industry is being forced to move away from the use of lead due to its toxicity [19]. Some lead-free glasses with low T _g and sufficient chemical durability include Zn–borophosphate, Zn–Sn–borophosphate [15], SnO–ZnO–P₂O₅ [20, 21], and alkali aluminophosphate glasses [15]. Borate glasses may be preferred in solder glasses for various reasons. Advantages include low melting and application temperature, low viscosity at working temperatures, and low toxicity [22].

Optical and optoelectronic applications

Non-silicate glasses are preferred over silicate glasses in certain optical, optoelectronic, and photonic devices due to their unique properties. Mass production of specially formulated glasses based on borate, phosphate, vanadate, germanate, or tellurite systems is not feasible but custom made glass components are produced for specific applications and for research purposes in physics, electronics, astronomy, etc. Among such special glasses, borate glasses are of interest mainly for four reasons:

• their low melting temperatures,

• their compatibility with rare earth elements and transition metals,

• their wide glass-forming range, and

• second- and third-order optical nonlinearity.

Borate glasses are capable of vitrifying over a wide range of compositions and host rare earth oxides such as La₂O₃, Pr₆O₁₁, Nd₂O₃, and Sm₂O₃ [23]. One specific application of such glasses is Faraday rotators [24]. The Faraday effect or Faraday rotation is a magneto-optical effect. It involves the rotation of the polarization plane of linearly polarized light under the influence of a magnetic field. The angle of rotation is expressed by

$$ \theta = VdH, $$

(1)

where θ is the angle of rotation (rad), V is the Verdet constant (rad/T-m), d is the length of the light path through the magnetic field (m), and H is the magnetic field strength (T). Depending on the ions incorporated in them, glasses typically exhibit either diamagnetic or paramagnetic behavior. Diamagnetic glasses commonly have a small positive Verdet constant with a small dependence on temperature. On the other hand, paramagnetic glasses have large, negative Verdet constants which are generally inversely proportional to temperature. According to the specific need in a device, diamagnetic or paramagnetic glasses may be preferred. However, paramagnetic glasses that contain a high concentration of rare earth ions exhibit large Verdet constants and large regions of transparency in the spectral range involved, which broadens the range of applications. Most optical silica glasses are diamagnetic and exhibit Verdet constants between 4 and 30 rad/(T-m) measured at 633 nm [25]. Although some special silica and other diamagnetic glasses have relatively high Verdet constants, they may cause self-focusing of an intense laser beam, which is not suitable for a high power laser system [26]. On the other hand, rare earth ion containing glasses with low refractive indices may solve this problem. Furthermore, the Faraday effect can be increased by lowering the temperature of the medium, which would increase the possibilities. Verdet constants of −71 and −30 rad/(T-m) have been reported at room temperature for Tb³⁺-doped (Hoya FR-5) and Ce³⁺-doped (Hoya FR-4) borate glasses, respectively [27]. A comparison of Tb³⁺-doped borate, fluoride, and fluorophosphate glasses indicated that the borate glass exhibited the largest Verdet constant [26].

Nonlinear optics is a growing field of science and technology where special glasses and other transparent materials are needed. Applications of nonlinear optical phenomena include frequency generation, optical communication, optical switching, and optical computing [28].

In linear optics, the refractive index of a medium remains constant and polarization of light obeys a linear dependence to an external optical electric field E [27],

$$ P = \varepsilon{_{0}}\chi^{(1)} E, $$

(2)

where ε ₀ is the permittivity of vacuum and χ ⁽¹⁾ is the linear susceptibility, expressed by,

$$ \chi^{(1)} = n^{2} - 1, $$

(3)

where n is the refractive index of the medium in which light travels. Glasses which exhibit linear susceptibility have passive uses such as optical fibers. However, some glasses have been shown to exhibit optical nonlinearity at very high light intensities typically involved with high intensity lasers. This effect is mostly observed in glasses containing highly polarizable ions and exhibiting high refractive index [25]. Non-linear polarization is expressed by a power series [27];

$$ P = \varepsilon_{0} \left( {\chi^{(1)} E + \chi^{(2)} E^{2} + \chi^{(3)} E^{3} + \ldots } \right), $$

(4)

where χ ⁽²⁾ and χ ⁽³⁾ signify second- and third-order susceptibility, respectively. Second- and third-order nonlinearity in glasses have potential applications such as planar waveguides, active optical fibers, photonic modulators, optical data storage, telecommunication, spectral extension of laser sources, and optical switches [27, 29, 30].

The refractive index of materials that exhibit third-order susceptibility is described by,

$$ n = n_{0} + n_{2} I, $$

(5)

where n ₀ is the linear refractive index, n ₂ is the intensity-dependent refractive index, and I is the laser intensity [31]. The relationship between n ₂ and χ ⁽³⁾ is given by the following equation:

$$ n_{2} = 5.26 \times 10^{ - 6} \chi^{(3)} /n_{0}^{2} . $$

(6)

Here, units of χ ⁽³⁾ and n ₂ are esu and m²/W, respectively [28].

Even-order optical nonlinearities are forbidden in centro-symmetric materials such as glass. Therefore, second-order nonlinearities are normally not observed in glass. However, in glasses where the macroscopic inversion symmetry is broken, second-order nonlinearity can be observed. The χ ⁽²⁾ value of thermally and electrically poled fused silica was reported to be as high as 1 pm/V [32]. A recent study showed that second-order nonlinearity in bismuth zinc borate (BZH) glass is limited due to electronic conduction [33]. With the aid of a model which takes into account both electronic and ionic conduction, the χ ⁽²⁾ value was increased from ~0.5 to ~2 pm/V. Efficient thermal poling could be achieved by using a 3 μm BZH film waveguide on a 150 μm thick glass substrate configuration.

The linear refractive index and third-order nonlinearity in borophosphate as well as La₂O–MgO–B₂O₃ glasses were observed to increase with Ti or Nb concentration. χ ⁽³⁾ values as high as 55 × 10⁻²² m²/V² were detected in sodium borophosphate glasses [30]. A comparative study involving praseodymium, neodymium, samarium, and lanthanum borate glasses showed that the best χ ⁽³⁾ values were obtained in praseodymium borate glasses, ranging from 2.1 × 10⁻²⁰ to 2.6 × 10⁻²⁰ m²/V², which are about 60 times larger than that measured for silica glass [23].

Major material selection criteria in nonlinear optics include high nonlinear susceptibility, high transmittance at all wavelengths involved, high resistance to laser damage, high chemical stability, and fast temporal response [31].

Various borate glasses have been studied in order to determine their nonlinear optical properties and to develop suitable glasses for specific applications. Optically stimulated luminescence (OSL) is a method that has been in use since the 1980s and its popularity has been growing ever since. Applications include archeological and geological dating through dose determination in natural materials such as quartz and feldspar, personal radiation dosimetry, and radiation diagnostic imaging [34]. Barium aluminoborate glass was considered as a candidate material for OSL dosimetry because it yields an intense luminescence when stimulated by blue light. A second OSL signal is stimulated by IR, although this is not as intense as the blue light stimulated signal. The OSL signal was reported to show significant fading (30 % in 1000 s) which was considered to be a serious drawback for dosimeter applications. However, there is the possibility of its employment in real time dosimetry [35].

Significant effort has been dedicated to understand the effect of transition metal and rare earth ions on the structure and optical properties of borate glasses. Many glass matrices doped with rare earth ions were demonstrated to operate as efficient luminescent materials and solid state lasers in the visible and near-IR regions. Lead borate and bismuth borate glasses doped with rare earth ions have received considerable attention due to their potential use in laser technology. Pisarska, Pisarski, and coworkers studied the luminescence behavior of rare earth doped lead borate glasses [36–38]. Two characteristic bands were found in the luminescence spectra of Dy-doped lead borate glass; one at 480 nm (blue) due to ⁴F_9/2–⁶H_15/2 and the second one at 573 nm (yellow) due to ⁴F_9/2–⁶H_13/2 of Dy³⁺ ions. The luminescence bands shifted to longer wavelengths and the yellow to blue luminescence ratios (Y/B) increased with increasing PbO:B₂O₃ mass ratio in (90−x)B₂O₃–xPbO-–Al₂O₃–3WO₃–1Dy₂O₃ (x = 30, 45, 60, 72, 75, and 80 wt%) except for the 18B₂O₃–72PbO–6Al₂O₃–3WO₃–1Dy₂O₃ system (represented by 1:4). A higher Y/B ratio is usually preferred because the intense yellow emission can be used as a visible solid state laser and raising this ratio also helps to obtain a single line instead of two (blue and yellow) lines [39]. Alternatively, an appropriate combination of these two lines can be used to generate white light by the glass [40]. Increasing the activator (Dy³⁺) concentration from 0.5 to 3 wt% in the 1:4 system increased the Y/B ratio from 1.05 to 1.15 while reducing the luminescence lifetime from 0.47 to 0.29 ms. Similarly, substitution of PbO by 9 wt% lead halides, PbX₂ (X = F, Cl, Br), in the 1:4 system increased the Y/B ratio from 1.08 to 1.32 while the luminescence lifetime decreased from 0.47 to 0.40 ms [36].

Another Dy³⁺ doped borate system studied in detail with regard to yellow/blue luminescence is the alkali borate glass system (95−x)B₂O₃–xNa₂O–5CaO–0.5Dy₂O₃. Increasing the Na₂O content from 5 to 30 mol% in this system increased the Y/B ratio from 1.05 to 1.5. Increasing the Na₂O content from 10 to 30 mol% decreased the luminescence lifetime from 0.85 ms to 0.54 ms. A strong correlation was found between the Y/B ratio and Ω₂/Ω₆ ratio, where Ω_t represent Judd–Ofelt parameters [39].

Bi₂O₃ is a heavy metal oxide similar to PbO that has been employed to formulate glass systems with useful nonlinear optical properties. Binary Bi₂O₃–B₂O₃ glass compositions have been studied in detail with regard to their third-order nonlinearity. The large third-order nonlinearity in these glasses was attributed to [BiO₄]⁵⁻ groups that possess significant hyperpolarization capability.

The effect of TiO₂ addition on nonlinear optical properties in Bi₂O₃–B₂O₃ glass was investigated. It was observed that third-order nonlinearity of such glasses increases with increasing TiO₂ concentration. The highest χ ⁽³⁾ value obtained was 3.8 × 10⁻²⁰ m²/V². The main source of nonlinear absorption at 800 nm laser wavelengths was determined to be two-photon absorption [41].

Chen et al. [42] investigated the spectroscopic properties of Bi₂O₃–B₂O₃ glasses doped with Nd³⁺. The maximum radiative quantum efficiency, stimulated emission cross section, and luminescence lifetime were determined as 54 %, 3.9 × 10⁻²⁰ cm², and 0.095 ms, respectively. Except for the luminescence lifetime, the spectroscopic properties are better than various commercial silicate laser glasses. The same team also studied the spectroscopic properties of Bi₂O₃–B₂O₃ glasses doped with Er³⁺ [43]. In this glass system, the best luminescence lifetime was measured as 0.72 ms while the maximum quantum efficiency and stimulated emission cross section were measured as 25 % and 0.94 × 10⁻²⁰ cm².

Raju et al. [44] studied the optical absorption and NIR emission properties of CdO–Bi₂O₃–B₂O₃ glasses doped with Nb³⁺. Ω₂ values ranged between 4.893 × 10⁻²⁰ and 4.248 × 10⁻²⁰, which were reported to be significantly larger compared to other glass hosts. Glasses excited by 514.5 nm (Ar⁺) laser resulted in NIR emission bands at 900, 1069, and 1338 nm. These bands were attributed to transitions (⁴F_3/2–⁴I_9/2), (⁴F_3/2–⁴I_11/2), and (⁴F_3/2–⁴I_13/2), respectively. The transition at 1069 nm was the one with high intensity. The emission intensity increased as the amount of Nd³⁺ increased from 0.5 to 1.5 mol% but decreased at higher fractions. The stimulated cross section values, calculated as 8.3–15.5 × 10⁻²⁰ cm² for the transition at 1069 nm, were much larger than those reported for commercial laser glasses (ranging between 1.9 and 4.8 × 10⁻²⁰ cm²). These studies show that lead and/or bismuth borate glasses doped with rare earth ions such as Dy³⁺ and Nd³⁺ yield candidate glasses for laser applications.

In addition to their role in nonlinear optical properties, it was shown that trivalent transition metal ions (Cr³⁺) and rare earth ions (Eu³⁺ and Dy³⁺) can be used as spectroscopic probes for studying the structure and local symmetry of lead borate glasses with different PbO/B₂O₃ ratios [45].

Biomedical applications

The bioactivity of some and the low chemical durability of most borate glasses open up some interesting possibilities in the biomedical field. Bioactive glasses have been used as implant materials for various bone parts and as fillers in dental applications since the 1960s [46]. The prototype material for such applications is known as bioglass or 45S5 which is based on silica as the glass former. Bioactive glasses attach to the bone chemically and form a strong bond. The formation of a hydroxycarbonate apatite (HCA) or hydroxyapatite (HAp) layer on glass implants or fillers plays a critical role in these applications. Like 45S5 and other silica-based compositions, some borate glasses developed for biomedical applications also exhibit the formation of HCA or HAp at the glass/tissue interface. One example is 13-93B3 glass, which is a Na–K–Ca-borate glass with no silica [47]. The bioactivity of Na₂O–CaO–B₂O₃ glasses with three different compositions was confirmed with XRD; FTIR and SEM analysis. SEM analysis indicated that an interconnected HA network of nearly spherical particles formed on the surface of the glass samples immersed in an aqueous phosphate solution for 3 days. As the molar ratio CaO:B₂O₃ increased, the average size of HA particles decreased, ranging between approximately 6 and 2 µm. However, higher magnification images revealed that these particles actually consist of finer crystals of size 30–70 nm [48].

Although borate glasses were shown to exhibit good compatibility with tissues and support cell proliferation, the high amount of boron ions released during dissolution may become toxic. Nevertheless, it has been reported that small concentrations of boron have an important role in bone formation, maintenance, and regeneration [48, 49]. In order to optimize the dissolution and boron release rate, borosilicate (13-93B1) and strontium borosilicate (2B–6Sr, 2B–12Sr) compositions have been developed. Table 2 shows the composition of various borate glasses considered for biomedical applications. The incorporation of strontium to borate glass was shown to decrease the cytotoxicity significantly [50]. Several in vivo animal studies with bioactive borate glasses reported that no adverse or toxic effects have been observed in the kidney, liver, or blood of the animals [51]. Glasses in the Na₂O–CaO–B₂O₃ system (B18, H12) have been specifically designed for titanium alloys since borate glasses exhibit much better bonding to titanium compared to silica glasses [52, 53]. Some evidence of the cytocompatibility of this family of glass was demonstrated by in vitro cell culture with human mesenchymal stem cells and osteoblasts derived from such cells [54].

Table 2 Borate glass compositions developed for biomedical applications

Many borate glasses possess biodegradable (or bioresorbable) properties rather than bioactive ones. Namely, they tend to dissolve in aqueous media. Some display both biodegradable and bioactive properties. Biodegradable glasses offer potential use as degradable temporary implants since they dissolve in a controlled manner in the body with the help of the surrounding environment and they are eventually replaced by the desired regenerated tissue [55]. H6b glass exhibited biodegradable behavior in simulated body fluid where the amount of dissolution showed a linear dependence on time [56] while some HAp formation was also observed at the surface. This is a glass based on the Na₂O–CaO–B₂O₃ system, which was developed for biomedical applications [57].

Another potential application of biodegradable glass is controlled drug delivery. Materials presently used for implanted drug delivery systems for the treatment of bone infection or bone regeneration present important disadvantages. Poly(methyl methacrylate) which has been in clinical use for a long time requires additional surgery to remove the inert part after serving as a drug delivery vehicle. Some polymers such as poly(lactic acid) and collagen are biodegradable but they don’t form a chemical bond with bone. The degradation of bioactive silica-based glass requires up to 1 year in the body in bone regeneration applications. A calcium borate glass scaffold was developed to serve as a device for drug release in the treatment of bone infection. 90 % of the glass in the device was dissolved within a week and HAp formation was observed. Drug release was completed in 4 days. In vitro and in vivo studies showed that such borate glasses are promising candidates for use as a scaffold and local drug delivery device for treating bone infection [58].

Bi et al. [51] studied the effect of bioactive borate glass scaffold morphology and composition on bone regeneration, angiogenesis, and conversion to HA. Glass compositions designated as 13-93B3 (see Table 2) and 13-93B3–Cu (same glass doped with 0.4 wt%CuO) were used. Copper ions are known to promote angiogenesis (formation of blood vessels) and the proliferation of endothelial cells, which are important processes for wound healing. Trabecular scaffolds with a microstructure similar to human trabecular bone have been prepared using a polymer foam replication technique. Oriented scaffolds with columnar pores were prepared by the unidirectional freezing technique. Figure 2 shows the morphologies of trabecular, oriented, and fibrous scaffolds employed. Defects of 4 mm in diameter were created in rat parietal bones. These defects were implanted with bioactive glass scaffolds, which fully converted to HA within 12 weeks. Micro-CT analysis of calvarial defects was conducted after 12 weeks following implantation. Figure 3 shows micro-CT images of bioactive glass implants of different morphologies and compositions. A higher fraction of bone growth was observed in scaffolds with trabecular microstructure. Doping the 13-93B3 glass with copper resulted in a significant increase of new bone growth only in the fibrous scaffolds (15 vs. 33 %, respectively). The amount of blood vessel formation was determined from stained calvarial sections (Fig. 4). The average blood vessel concentration in the cross sections was measured as 4.0 and 4.3 % for 13-93B3 and 13-93B3–Cu trabecular scaffolds, respectively, while it ranged between 2.7 and 3.3 % for other scaffold morphologies. It was also observed that borate glass scaffolds performed better than silicate 45S5 particles used as positive control in this study.

Fig. 2

Morphologies of trabecular (a, b), oriented (c, d), and fibrous (e, f) bioactive glass scaffolds. Reprinted from Ref. [51] with permission from Elsevier

Fig. 3

Micro-CT images of 13-93B3 and 13-93B3–Cu bioactive glass scaffolds with different morphologies and 45S5 particles, 12 weeks after implantation to rat calvarial defects. Reprinted from Ref. [51] with permission from Elsevier

Fig. 4

Micrographs of PAS stained sections of calvarial defects in rats implanted with bioactive glass, 12 weeks after implantation. a Trabecular 13-93B3 scaffold; b trabecular 13-93B3–Cu scaffold; c oriented 13-93B3 scaffold; d oriented 13-93B3–Cu scaffold, e fibrous 13-93B3 scaffold; f fibrous 13-93B3–Cu scaffold; g 45S5 particles as positive control. White areas marked with (asterisk) are remnants of glass scaffolds or particles. Arrows indicate blood vessel formation with red blood cells colored green. Scale bar 100 µm. Reprinted from Ref. [51] with permission from Elsevier (Color figure online)

Borate glass delivery systems composed of 6Na₂O–8K₂O–8MgO–22CaO–54B₂O₃–2P₂O₅ glass pellets infiltrated with the antibiotic vancomycin were used for the treatment of chronic osteomyelitis in rabbits. The borate glass delivery system was found to be effective for such a treatment [59]. Borate glasses are becoming important in carrier and other biomedical applications because of the important role of boron during embryogenesis and bone growth [60].

Recently a wound healing product was developed by Mo-Sci, a company specialized in precision glass technology. The product with the trade name DermaFuse is composed of borate based glass nanofibers and supplied in the form of bandage-sized pads. It is aimed for treatment of difficult to heal wounds such as venous ulcers experienced by diabetics and pressure ulcers (bedsores) found in bedridden patients or individuals confined to a wheelchair. Initial tests with more than 60 human subjects indicated that the product is successful in healing deep, complicated, and long-standing wounds. The composition of nanofibers are given as 51.5–53 % B₂O₃, 20 % CaO, 6 % Na₂O, 12 % K₂O, 4 % P₂O₅, 5 % MgO, 0–0.4 % CuO, and 0–1 % ZnO (in wt%). Calcium is believed to be an important factor in the healing mechanism, especially in the later stages, because it helps the migration of epidermal cells. The nanofibers create a clot-like structure and provide an antimicrobial environment. Furthermore, it is believed that as the glass is dissolved, various ions, such as Cu²⁺, are absorbed at the wound site and they stimulate innate cellular and genetic healing processes. [47, 61, 62].

Another borate based glass that is of interest for the biomedical field is antimicrobial glass, developed by the same company. This is a silver containing glass powder with a composition of 40–80 % B₂O₃, 0–70 % P₂O₅, 0–30 %Al₂O₃, 0–12 %Li₂O, Na₂O, K₂O, and 1–30 %Ag₂O (wt%). The antimicrobial effect is provided by silver ions dissolving into aqueous liquids at a rate that is controlled by composition and particle size [62]. One of the biomedical applications of such powders is antimicrobial additives for teeth restoration materials [63].

Calcium sodium borate glass microspheres, 45–90 µm in diameter, produced by flame spheroidization, were transformed into HA as a candidate material for protein separation applications. HA is commonly used in protein separation and purification. The initial spherical shape of glass particles was preserved during transformation. Experiments with bovine serum albumin, chicken egg lysozyme, and horse heart myoglobin indicated that similar protein separation performance could be attained compared to much finer commercial spherical HA particles produced by more complex processes. Spherical particles performed better than irregular particles made of the same material [64].

Radioactive waste immobilization and gamma-ray shielding

A limited number of studies investigated the possibility of using borate glasses as a matrix for high level radioactive waste (HLW) [65–67]. This is a very challenging application since one of the main requirements for such an application is centuries-long integrity and durability of the final product. Although borosilicate glasses are commonly used for nuclear waste immobilization practices, at the absence of silica, borate glasses have very low chemical durability in general, as will be discussed in detail under the section discussing this property. One of the first studies related to nuclear applications of silica-free borates investigated the dissolution rate of CaBAl glass with additions of U₃O₈, Cs₂O, and SrO in water at 70 °C [65] While CaBAl has relatively high chemical durability, the additions of waste oxides reduced the chemical durability significantly. It was suggested that the formation of a Ca-aluminate or Ca(OH)₂ gel layer on the glass surface stabilizes the dissolution process. The effect of γ-irradiation on chemical durability was also studied. It was shown that as the dose of γ-irradiation increases, the weight loss increases remarkably during chemical durability tests. Remizov et al. [66] investigated borophosphate glasses as potential materials for HLW immobilization. Boro-organic reagents were used to immobilize liquid HLW in borophosphate glass. However, no detail was given on the dissolution behavior or other properties of the resulting glass.

A recent study reported that lead borate glasses offer some advantages over borosilicate glasses for HLW immobilization applications [67]. Two of these advantages are the low melting points (T _m < 550 °C compared to T _m = 1000–1600 °C) and lower viscosities compared to borosilicate glasses. It was demonstrated that lead borate glasses with PbO content over 70 mol% demonstrated very limited dissolution. Simulated waste glass samples containing SrO or Cs₂O indicated that an alteration layer seems to inhibit the dissolution of the former while this was not the case in Cs₂O-containing lead borate glasses.

Singh et al. [68] investigated the possibility of using lead borate and bismuth lead borate glasses as a shield for γ-irradiation. The radiation shielding properties of these glasses were shown to be significantly better than various types of concrete at 511 and 662 keV line energies. The mass attenuation coefficient increased almost linearly from about 0.11 to 0.13 cm²/g as the mole fraction of PbO increased from 0.2 to 0.5 in lead borate glasses. Shielding toward gamma irradiation was also observed in bismuth borate glasses (60Bi₂O₃–40B₂O₃) especially in the visible region. This was attributed to high density and large atomic number of Bi³⁺ ions and their ability to retard the effects of gamma-rays [69]. Kaur and Singh [70] studied the gamma-ray shielding properties of lead borate glasses doped with 0.1 mol% Al₂O₃ at a photon energy of 662 keV. Increasing the PbO mole fraction from 0.25 to 0.45 in this system increased the mass attenuation from ~0.092 to ~0.096 cm²/g. Ruengsri [71] calculated the theoretical gamma-ray shielding properties of borate, silicate, and phosphate glasses containing PbO and compared them to two types of standard shielding concrete. Figure 5 compares the mass attenuation coefficients of three types of glass with varying PbO concentrations and two types of shielding concrete. As seen from this figure, the radiation shielding performance of PbO-containing glasses does not change significantly with respect to the matrix glass. However, all glass materials exhibit better shielding when compared to standard shielding concrete. Thus, PbO and/or Bi₂O₃-containing borate glasses are candidate radiation shielding materials with the advantages of transparency and smaller thickness than concrete shielding components.

Fig. 5

Increase of mass attenuation coefficient with increasing amounts of PbO in borate, silicate, and phosphateglass at 662 keV photon energy. A comparison is shown with two types of shielding concrete. Reprinted from Ref. [71] (Color figure online)

Glass-forming ability, glass stability, and crystallization

Obviously the first requirement in applications where glass is a preferable material is the ability to form a glassy (or amorphous) structure in the desired composition. According to the kinetic theory of glass formation, any material can be transformed into glass [1]. The main limitation is the cooling rate from the molten state. The concept of glass-forming ability (GFA) is directly connected to nucleation and crystallization. Shelby defines GFA as “the critical cooling rate required to prevent crystallization of a specified volume fraction of the sample” [1]. This definition is vague because it doesn’t tell us what the “specified volume fraction” is. Usually this fraction is related to the technical ability to detect crystals in molten glass and ranges from 1 % to 1 ppm by volume [72, 73]. If a liquid is cooled from the melting (liquidus) temperature, T _m, at a rate higher than the critical cooling rate, R _c, the liquid can form a glass [74]. The slope of the cooling line tangent to the nose of the time–temperature–transformation (TTT) curve (line 2 in Fig. 6) gives the R _c value. A faster cooling rate than R _c will yield a transparent glass (line 1). If the cooling rate is slower than R _c, crystallization will occur (line 3). A system where nucleation and growth of crystals is slower, i.e., where glass formation is easier, yields a smaller R _c value. Thus, the smaller the critical cooling rate of a glass, the better its GFA. Although R _c is a very useful indicator of GFA, it is not readily available for many oxide systems. Calculation of R _c requires experimental TTT diagrams or crystal nucleation and growth rates in a wide temperature region, which are rarely available probably because related measurements are very time consuming [1, 75]. Factors which favor high GFA include a high barrier to nucleation, a high viscosity at the nose temperature of the TTT diagram, and complex melt structures or compositions, since they inhibit the rearrangement of the melt into a crystalline structure [1].

Fig. 6

TTT diagram for a hypothetical glass-forming composition

A more useful method, although indirect, is to use thermal analysis data to estimate GFA. The stability of melts against crystallization has been commonly termed glass stability (GS). The same term is also used to describe the resistance of a glass (below the crystallization temperature) against crystallization upon reheating. Nascimento et al. [72] compared critical cooling rates of eight stoichiometric glass-forming oxides with seven different GS parameters. Six of these glass systems present surface crystallization and two of them bulk crystallization. The study demonstrated that most of the GS parameters that consist of three characteristic DTA/DSC temperatures show excellent correlation with GFA while those which consist of only two characteristic temperatures (T _g and T _m) do not show a good correlation. Comparison of critical cooling rates and GS parameters in these glasses showed an approximately inverse linear relationship between ln(R _c) and GS parameter K. According to this relationship, as K increases, ln(R _c) decreases. In other words, as the GS parameter increases, GFA also increases.

Among the six GS parameters tested, three of them, namely the parameters of Weinberg (K _W), Hrubÿ (K _H), and Lu and Liu (K _LL) provided the best correlation factors, which were in the range of 0.95 to 0.97. These three parameters are given in Eqs. (7–9), respectively [72].

$$ K_{W} = \frac{{T_{p} - T_{g} }}{{T_{m} }} $$

(7)

$$ K_{H} = \frac{{T_{x} - T_{g} }}{{T_{m} - T_{x} }} $$

(8)

$$ K_{LL} = \frac{{T_{x} }}{{T_{g} + T_{m} }}, $$

(9)

where T _p represents the maximum of the relevant crystallization peak, T _x is the onset temperature for crystallization, and T _g is the glass transition temperature. Using the three parameters given in Eqs. (7–9) and data available in the literature (Table 3), a comparison of GFA and GS in different binary borate glasses has been made (Table 4). Based on these results, it can be inferred that 17Li₂O·83B₂O₃ has the highest GFA and the highest stability against crystallization among the Li₂O–B₂O₃ compositions listed in Table 4. For the PbO–B₂O₃ and Bi₂O₃–B₂O₃ systems, the highest GFA and GS values are expected in glasses with 40 mol% PbO and 25 mol% Bi₂O₃, respectively. The highest GS parameters for three systems listed in Table 4 correspond to the same composition within each system (shown with bold numbers), suggesting a good correlation between different GS parameters. Comparison of the three different systems indicates that, in general, PbO–B₂O₃ glasses have higher GS values, with respect to Li₂O–B₂O₃ and Bi₂O₃–B₂O₃ glasses. This also implies that PbO–B₂O₃ glasses are the easiest to form among these three binary systems.

Table 3 Some important properties of binary borate systems

Table 4 Glass stability parameters calculated for various borate glasses

Phase separation

Even if crystallization can be avoided during or after glass formation, many glass-forming systems are inclined to phase separate [76–78]. Table 5 lists the compositional ranges for glass formation and phase separation for various binary borate systems. Some of the data such as glass formation range and immiscibility range may seem to be conflicting in this table. The reason for this is the different sources used to create it. Glass formation ranges indicate compositions where homogenous clear glass can be obtained. This normally excludes immiscibility ranges where mixed glasses can be obtained by quenching from a region with two immiscible liquids. However, sophisticated techniques are required to be sure whether immiscibility occurs. It has been demonstrated that glasses which appear to be perfectly clear may possess a phase separated structure at the nanometer scale [25]. Phase separation is not necessarily a phenomenon that prevents a glass from being used in an industrial application, but in some cases it may have an unpredictable or undesirable impact on the targeted property. For example, if the phase separated structure has an interconnected morphology instead of isolated spheres in a matrix, and if one of the interconnected phases has a significantly inferior chemical durability, the overall chemical durability will, to a great extent, be determined by the phase with inferior durability. The mechanical properties can also be greatly impaired. In the case of an interconnected morphology, the weaker phase will form the weakest link for fracture and determine the overall strength. If isolated phases are present, a rule of mixture could be used to roughly estimate mechanical properties such as strength, hardness, and elastic modulus [79]. More precise estimations should be based on the Hashin–Shtrikman/Kerner relations since these were shown to best fit phase separated glasses such as sodium silicate and alkali borate glasses [80].

Table 5 Ranges of glass formation and phase separation in binary borate systems

The first thing to consider regarding phase separation in certain borate systems is the phase diagram, if available, and work with compositions which allow glass formation without phase separation. For binary borates, another alternative to consider is to convert a two-phase melt into a single-phase by adding a small amount of a third oxide [18, 81]. Usually, alkali oxides are recommended as the third phase for binary RO–B₂O₃ systems since they act as oxygen donors. For example, more than 3 mol% K₂O is known to eliminate immiscibility in CaO–B₂O₃, SrO–B₂O₃, and BaO–B₂O₃ systems [81].

There are two separate glass formation ranges in some alkali borate systems including Na₂O–B₂O₃, Rb₂O–B₂O₃, and Cs₂O–B₂O₃ (Table 5). The first range (or the only range in the case of Li₂O–B₂O₃ and K₂O–B₂O₃) lies between 0 and roughly 50 mol% alkali oxide while the second range is more limited. In the case of alkaline earth borates, the glass formation ranges are usually narrower. This is partly explained by the high melting points of alkaline earth oxides such as BaO and MgO, which raise the liquidus temperature rapidly with increasing alkaline earth oxide content. The glass formation range is very broad in the case of PbO–B₂O₃ and Bi₂O₃–B₂O₃ systems. Furthermore, liquidus temperatures are quite low (Table 3). Both of these conditions make it convenient to make glass in these systems. In the case of alkali borate systems, on the other hand, liquidus temperatures rise rapidly with increasing alkali oxide content. So, most of the alkali borate compositions have melting points similar to silicate or borosilicate glasses [81].

Two possible routes to phase separation in glasses are stable and metastable phase separation. Stable phase separation will occur if the phase diagram of the glass-forming system contains a region of liquid–liquid immiscibility above the liquidus temperature and if the glass composition corresponds to this region. Cooling from this region will result in two separate phases in the final product. Many binary borate systems exhibit such stable phase separation, including Bi₂O₃–B₂O₃ [82], BaO–B₂O₃ [83], Nb₂O₅–B₂O₃ [84] and La₂O₃–B₂O₃ [85]. The phase equilibrium diagram for the system BaO–B₂O₃ is shown in Fig. 7 as an example, indicating the region of stable immiscibility, the upper consolute (critical) temperature, T _c, and corresponding critical composition, θ _c. Interestingly, in most such systems, the region of immiscibility occurs in the B₂O₃ rich side of binary phase diagrams. The PbO–B₂O₃ system also exhibits stable phase separation [86]. The immiscibility region lies between 0.4 and 20 mol% PbO [87]. This is one of the commonly studied systems regarding immiscibility because the phase separated regions have a big difference in terms of density, which facilitates microstructural analysis. Zhu et al. [88] studied the microstructure of phase separated lead borate glasses prepared under low gravity. Figure 8 shows the resulting microstructure in 3.5PbO–96.5B₂O₃ glass. The microstructure is composed of a continuous borate-rich phase in which lead-rich droplets are distributed. It was reported that the addition of a small amount (e.g., 4 mol%) of Al₂O₃ reduces phase separation in the PbO–B₂O₃ system [89].

Fig. 7

Phase diagram of the binary system BaO–B₂O₃ (Adapted from Ref. [114])

Fig. 8

a SEM micrograph of phase separated 3.5PbO·96.5B₂O₃ glass and b detail of lead-rich phase (Courtesy of Dongmei Zhu, State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an, Shaanxi 710072, People’s Republic of China)

The second route for phase separation in glasses is metastable, which occurs below the liquidus temperature. Figure 9 shows a typical phase diagram with a region of metastable or subliquidus immiscibility. Such diagrams are useful to predict the type of microstructure (isolated droplets or interconnected phases) based on the composition and quenching rate [1]. Two binary borate systems known to exhibit metastable immiscibility are Li₂O–B₂O₃ [90] and B₂O₃–SiO₂ [91]. Although it was claimed in the 1960s that other alkali borate systems including Na₂O–, K₂O–, Rb₂O–, and Cs₂O–B₂O₃ also exhibit metastable immiscibility [90], later studies suggested that no miscibility gaps existed or no phase separation occurred in them [92, 93]. Yet electron microscopy imaging provided proof of phase separation at the nanoscale in sodium borate glasses [94]. Du et al. [95] found that it takes long time in 15Na₂O–86B₂O₃ glass to reach equilibrium during phase separation. At 500 °C, 20 h of heat treatment was necessary to fully convert glass samples to Na₂O·9B₂O₃ and 3Na₂O·B₂O₃. The long time for reaching equilibrium was attributed to the difficulty in crystallizing the Na-rich phase due to over-coordination of oxygen atoms during conversion of some trigonal borates to tetrahedral borates. Binary immiscibility diagrams are not well established and controversy remains in many borate or silicate systems. It is also questioned whether immiscibility in the B₂O₃–SiO₂ system exists [1]. Further study would be very valuable to clarify the question of immiscibility. New tools and techniques need to be applied for more reliable results.

Fig. 9

Hypothetical phase diagram with a region of subliquidus immiscibility

Although normally undesirable, phase separation has, in some cases, been used to obtain unique materials such as porous, anisotropic, and nonlinear optic glasses [96]. For example, in the Vycor^® process, a microporous glass with a composition of nearly pure SiO₂ is obtained by leaching the alkali borate phase from the interconnected silica-borate structure [94]. It is essential to know the phase separation rate or the nucleation and growth rates in the systems of interest for optimum results when making practical use of phase separation. Some experimental results exist for BaO–B₂O₃ and PbO–B₂O₃ systems in this regard [88, 94, 97].

Chemical durability

So far, binary borate, particularly alkali and alkaline earth borate glasses could find no practical application since their chemical durability is unacceptably low. Table 6 gives a better idea of how their chemical durability compare with those of other glasses. Industrially useful levels of chemical durability have been obtained in CaO-aluminoborate (CaBAl) glasses developed for sealing applications [98]. The chemical durability of CaBAl glasses in water at 70–90 °C is in the order of 10⁻⁶ g/(cm²min) [99, 100]. However, the melting point of CaBAl glasses are higher than many binary borate glasses; for example that of CaBAl17 (50CaO·30B₂O₃·20Al₂O₃) is 940 °C [100].

Table 6 Dissolution rates of various glasses in water

Studies in the PbO–B₂O₃ system showed that the chemical durability of glasses in this system is much better compared to alkali borate (Na₂O–B₂O₃) or alkaline earth borate (CaO–B₂O₃) glasses. The aqueous dissolution rate of lead borate glasses at 90 °C is five to six orders of magnitude smaller compared to Na₂O–B₂O₃ glasses (Table 6). The chemical durability of glasses in the PbO–B₂O₃ system increases as the PbO content increases. This increase is not linear but there seems to be some compositional ranges where the dissolution mechanism remains the same. In the first compositional range where PbO = 30–50 mol% and B₂O₃ = 70–50 mol%, the dissolution rate is 2–3 × 10⁻⁶ g/(cm²min) (Table 6). The aqueous dissolution rate of 60PbO.40B₂O₃ glass is ~2 × 10⁻⁷ g/(cm²min). Lead borate glasses with high (70–80 mol%) PbO content exhibit the lowest dissolution rate, namely ~3.6 × 10⁻⁸ g/(cm²min) in this system. The addition of 5 mol% PbO reduces the dissolution rate of 30Na₂O·70B₂O₃ by about two orders of magnitude. The relatively high chemical durability in borate glasses with 5–50 mol% can be explained by a high content of BO₄ groups in the glass network. On the other hand, the significantly high chemical durability in borate glasses with more than 50 mol% PbO can be explained by an increased covalency of Pb–O bonds where PbO acts a network former [4, 101].

The addition of small amounts of P₂O₅ to SrTiO₃-aluminoborate glasses resulted in aqueous dissolution rates of about two orders of magnitude lower than found in the original glasses. This unexpected improvement in chemical durability was attributed to the replacement of BO₃ units by chemically resistant BO₄ units [100]. Subsequent studies investigated whether P₂O₅ additions could improve the chemical durability of simpler borate glasses. However, such an effect could not be observed in binary systems. For example, the aqueous dissolution rate of 30Na₂O.70B₂O₃ at 90 °C did not decrease but rather increase from ~1×10⁻³ to 2 × 10⁻³ g/(cm² min) with 5–20 mol% P₂O₅ additions [102].

A comparative analysis of aqueous dissolution rates of various borate and other types of glass (Table 6) leads to the following comments:

• In general, simple alkali- or alkaline earth-borate glasses possess very low chemical durability.

• On the other hand, some complex borate glasses such as SrTiO₃–B₂O₃–P₂O₅ glasses have comparable chemical durabilities to borosilicate glasses.

• Lead borate glasses with high lead content possess comparable chemical durability to borosilicate glasses. The former glasses are simple to make and have very low melting temperatures. Unfortunately, they are not suitable for applications such as biomedical fields due to the toxicity of lead. However, they are good candidates for innovative applications where toxicity is not a major issue.

Factors related to borate melts and potential implications on glass properties

Commercial or experimental glass making requires close control of process parameters and environmental factors for optimized and consistent results. Some factors commonly known to influence the quality of glass products are melting temperature, cooling rate, batch volume, annealing schedule, purity of raw materials, gas evolution from raw materials, volatilization of components, and mixing of the melt [1].

One of the attractive features of borate glasses is their considerably lower melting and processing temperatures, in the absence of silica or other refractory oxides. The melting temperature depends on the composition and purity. Melting temperatures of binary alkali borate glasses range between ~500 and 920 °C (Table 3). Among various glass-forming xR₂O–yB₂O₃ compositions, the maximum melting temperatures and corresponding amounts of R₂O are 919 °C & 21 mol% for Li₂O; 854 °C & 26 mol% for K₂O, and 800 °C & 19 mol% for Na₂O [2]. Alkaline earth borate glasses have typically higher melting points, ranging from ~880 to 1150 °C. The melting temperatures for binary PbO–B₂O₃ glasses range between ~470 and 720 °C, while those of Bi₂O₃–B₂O₃ range between ~650 and 730 °C.

In the glass transition region, addition of alkali oxides increases the viscosity until a maximum is reached at about 20–30 mol% R₂O. Further additions decrease the viscosity monotonically. Therefore, a viscosity-% R₂O curve creates a parabola with a maximum point. Viscosity curves for Sb₂O₃–B₂O₃ and Bi₂O₃–B₂O₃ melts exhibit very similar behavior to R₂O–B₂O₃ melts. In R₂O–B₂O₃ melt systems, the increase in viscosity up to the maximum has been interpreted as an increase of connectivity in the melt network through the destruction of boroxol rings and formation of BO_4/2R tetrahedra. Above the maximum, increased fluidity occurs due to non-bridging oxygen atoms in BO _n/2(OR)_(3−n) triangles (n = 1,2,3) [103]. According to these observations, glasses in alkali oxide-borate systems should be selected sufficiently far from 20 to 30 mol% alkali oxide compositions in order to obtain melts with lower viscosity for practical reasons. If, for any reason, 20–30 mol% alkali oxide is preferable in a borate glass, higher temperatures (T > 1000 °C) could be used for casting.

In the case of B₂O₃–SiO₂ and B₂O₃–GeO₂ systems, the transition range viscosity increases monotonically with increasing SiO₂ or GeO₂ content. The monotonic increase has been explained by the splitting of boroxol rings into individual BO_3/2 triangles which are linked to SiO_4/2 (Q4) or GeO_4/2 tetrahedra [1, 103].

Gases released from decomposition of carbonates, nitrates, and sulfates, or from the evaporation of moisture absorbed on particles can lead to bubbles. Small amounts of fining agents are added to the batch to reduce bubble formation. Even though most borate glasses have lower viscosities than silica glasses, bubble entrapment can still be a problem in them. As₂O₃ and Sb₂O₃ are the most efficient fining agents but they are toxic. Nitrates can be used instead of these toxic agents but they have low decomposition temperatures ranging from 500 to 800 °C. While they are not preferred for silica glasses, they may be suitable for low melting temperature borate glasses. CeO₂ is an acceptable substitute for As₂O₃ and Sb₂O₃ at higher melting temperatures.

Another potential problem in borate glasses is entrapped H₂O. Recent studies confirmed that the solubility of water is much greater in borate melts compared to silica melts. The solubility was measured as 5600–6500 ppm (mass/mass) of H₂O at 800 °C and 6400–6900 ppm at 900 °C in alkali borate melts, whereas the highest solubility for comparable silica melts is about 2300 ppm. The solubility of water in alkali borate melts was shown to decrease by a factor of two or more as the alkali oxide concentration increased from 0 to 40 mol%. The large amount of water in borate glasses results in a wide spread of glass transformation temperature values [104, 105]. Water dissolved in the melt tends to react with oxide glasses and cause a replacement of bridging oxygen atoms by hydroxyl groups. Such reactions result in weaker network connectivity, lower glass transformation temperatures, lower viscosities, and higher crystallization rates, in general [105].

Volatilization of borate species from glass melts is an important phenomenon to consider in glass making or high-temperature applications such as sealing. Volatile species responsible for weight loss were determined to be mainly BO₂ (g) in dry conditions and B₃H₃O₆ (g) in wet, reducing conditions [106]. Major (up to 20 %) weight losses have been found to occur at melting temperatures or above when borate was the only network former in the glass composition [107]. Studies with B₂O₃-containing glasses indicated that the amount of boron volatilization increases with the melting temperature, partial water pressure in the melting environment, and B₂O₃ fraction in the glass [106, 108]. Volatilization of borate species has to be kept under control because it would affect all glass properties. For example, the density and refractive index of E-glass increased as the boron content of glasses decreased [108].

Khanna et al. demonstrated that the properties of lead borate and bismuth borate glasses can vary significantly with melt annealing [109, 110]. The density and elastic modulus decreased while the glass transition temperature increased with increasing melt aging time (i.e., the duration of annealing at a given temperature above liquidus) in a bismuth borate glass with the composition 40Bi₂O₃–60B₂O₃. Similar results were obtained in PbO–B₂O₃ of various compositions [109]. In the case of bismuth borate glass, increasing the aging time from 15 to 220 min decreased the density from 5.84 to 4.97 g/cm³ and the longitudinal modulus from 86.3 to 80.3 GPa, while it increased T _g from 455 to 474 °C, respectively [110]. No selective evaporation of constituents from the melt was detected. Property changes were attributed to structural rearrangement of the melt from a high density amorphous to a low density amorphous state due to the structural transformation,

$$ {\text{B}}{\emptyset }_{4}^{ - } \Rightarrow {\text{B}}{\emptyset }{\text{O}}^{ - } , $$

(10)

where $\emptyset$ represents a bridging oxygen while O⁻ refers to a non-bridging oxygen. Time related effects were also reported for lithium tetraborate and sodium tetraborate melts but the change in glass density was only in the order of 3 % [111, 112]. The melt aging effect may be one of the reasons for significant variations in properties of binary borate glasses with the same nominal composition (Table 3). It was suggested that various properties of heavy metal oxide borate glasses can be tailored by changing the aging time of the corresponding melt without altering the composition or quenching rate [110]. More detailed analysis of melt aging effects on different glass properties would be valuable both for scientific and industrial reasons.

Implications for future research and applications

Various non-silicate borate glasses have been developed for applications where silica chemistry is not suitable or where the borate network offers certain advantages over silica or other types of glasses. While the unique structure and chemistry of borate glasses offers important advantages and possibilities, certain deficiencies such as low chemical durability necessitate the formulation of carefully selected compositions. This review discussed binary borate glasses in general. Among binary borates, lead borate systems offer certain advantages such as good chemical durability and low processing temperatures. However, the toxicity of lead will gradually restrict the applications of lead borate glasses due to regulations and environmental concerns. The low toxicity of bismuth makes it a candidate for replacement of lead in low melting point solders [113]. Similarly, bismuth borate glasses may be preferable for sealing and other applications that require low melting points. Typically, ternary and more complex borate glasses are needed for advanced applications. The effort of choosing the right composition requires a good analysis of the scientific literature available on the subject. The literature and collected data for similar compositions (Table 3) indicate that there is considerable variability in reported experimental data. Reasons for such variability may include H₂O dissolved in the glass network in the form of hydroxyl groups, melt aging, and experimental errors inherent in related testing and measurement equipment. The appreciation and minimization of such factors is a necessity for better control of glass preparation and resulting properties.

Although research on borate glasses dates back to more than a century ago, some basic investigations such as the development of phase diagrams of binary systems may be worthwhile repeating with more sophisticated experimental tools and thermodynamic calculations since the reliability of some older data is now under debate. For example, as mentioned earlier in this review, regions of immiscibility for alkali borate systems are still not well defined. Another important contribution to the study of borate glasses would be related to chemical durability. Data on the chemical durability of borate glasses in different media (water, acidic, basic, etc.) are quite limited and difficult to compare. While some standard testing methods exist, they are relatively complicated. Development of simple standards for chemical durability testing seems to be an important necessity. The effect of various factors such as water content and melt aging on glass properties is another field where further research would be valuable. Relatively new approaches such as computerized simulation would be useful for predicting long-term effects such as leaching from waste glasses in underground storage areas or volatilization of borate from sealing glasses used at elevated temperatures.