This chapter introduces some fundamental chemical and physical properties of glasses to be taken into account when designing and fabricating products based on bioactive glasses to be implanted inside the human body. The main emphasis is to explain the constraints to be taken into account from the materials science and chemical engineering points of view. The ultimate goal is to deliver the basic principles of glass science to serve as a platform for the various disciplines ranging from materials science to molecular biology, biochemistry, medicine, etc. The vision is that the steadily increasing multi-disciplinary experience provides the crucial knowledge needed for developing implants and scaffolds for controlled predetermined performance in the target application.

At first sight, the inherent brittle nature of glass does not make it a feasible material for implantable medical devices. At the same time, glass has several useful properties which support its utilization as a biomaterial. What is a biomaterial? Most biomaterials based on glasses or ceramics are designed to improve human health and the quality of life by restoring the function of living tissue and organs in the body. The single most important factor for a biomaterial is that it is able to be in contact with tissues of the human body without causing an unacceptable degree of harm to that body, i.e. the material is biocompatible.¹  Recently, considerable research efforts have been directed to tailor highly porous tissue engineering scaffolds not only for bone tissue but also for emerging soft tissue applications. Detailed understanding of the nature and properties of glass provides a thorough basis for assessing its potential in prospective biomedical applications. In general, choice of a material for a particular application is based on its performance, properties, fabricability, and manufacturing costs.

Due to its amorphous structure glass possesses several features which make it an optimal material for manifold applications. Simplified, conventional inorganic glasses are homogeneous mixtures of oxides of alkalis, alkaline earths, aluminium, boron, silicon, etc. Most modern uses of these so-called soda lime glasses, including flat glass, hollowware glass and fibre glass, rely on the transparency to visible light combined with some other property, such as good mechanical strength, adequate chemical durability or electrical resistivity. Essentially, these commercial glass types are fabricated via the inexpensive melting route. They are easy to shape into various product forms at high temperatures when present as viscous liquids. In addition to the conventional glasses, specialty glasses possessing certain functional properties are essential in the modern everyday environment.

The functionality of the speciality glasses is often connected with the optical properties. Bioactive glasses used in contact with the living body as implants or as tissue engineering scaffolds belong to the family of speciality glasses. In contrast to most other applications, the transparency is not an essential property for the current applications of bioactive glass. Bioactivity of glasses may be defined in different ways, but common to all of these definitions is the requirement that the surface composition and morphology of the glasses change upon implantation. Simultaneously, the concentrations of the inorganic ions in the surrounding extracellular fluid change. Only glasses within certain limited composition range fulfil the requirements of bioactivity, i.e. show the desired interaction with the living tissue. Interestingly, whether it be transparency or the controlled surface reaction, the origin of the functionality of the glass is the same—the amorphous glass structure. As explained in Chapter 6 the amorphous glass structure enables adjusting the physical and chemical properties of glasses within certain limits merely by changing the constituent oxides or their ratios. This gives interesting possibilities to tailor the glass composition for various clinical applications.

Bioactive glasses consist of the same oxides as conventional soda lime glasses but in proportions which give rise to large differences between several properties of these types of glass. The most marked difference dictating the bioactivity is the chemical durability; the bioactive glasses dissolve in aqueous environments at markedly higher rates than the soda lime glasses used, for example, in containers or windows. For the glass to be bioactive the dissolution rate must be compatible with the cellular processes so that the dissolving glass supports and enhances tissue regeneration and growth, while we expect the traditional glasses to be inert in their normal applications. Bioactive glasses thus provide temporary support to tissue healing and regeneration. Brittleness is one of the biggest challenges for utilization of bioactive glasses, especially in load-bearing applications. In future the problems with the brittleness may be solved by using the bioactive glass in composites together with polymers, or developing tissue engineering scaffolds with special architectures adapted to the requirements of loaded bone.

The first bioactive glasses studied as prosthetic materials were melt-derived compositions within the system Na₂O–CaO–P₂O₅–SiO₂.²  Glass was an interesting and smart choice for a material to be in contact with the human skeleton. The pioneering idea by Professor Hench, the inventor of bioactive glasses, was to develop a material that consists of elements abundant in the human body. In addition, the ratio between the oxides in the compositions studied was selected to favour rapid initial dissolution of alkalis from the glass surface in aqueous solutions followed by precipitation of an outer layer rich in calcium and phosphorus at the inner alkali-depleted silica layer. Virtually all glasses gradually dissolve in aqueous solutions, but the ion leaching rates and the tendency to form surface layers varies markedly depending on the total composition.³  It was hypothesized that if the composition of the calcium phosphate surface layer is similar to the hydrated calcium phosphate component in bone tissue, hydroxyapatite (HCA), the glass would not be rejected by the body.⁴  The compositions developing a HCA layer in vivo are today known as bioactive glasses.

Accordingly, the bioactive glasses were proven as advantageous materials in skeletal repair. Today, extensive research efforts are made to develop new compositions of bioactive glasses for bone and soft tissue engineering.⁵  Increasing knowledge of the influence of various ions released from the glasses on the tissue regenerative capability has inspired researchers to adjust and tailor bioactive glass compositions far beyond the original bone tissue related applications toward emerging areas in soft tissue regeneration. Though the history of bioactive glasses is longer than forty years, most commercial products today are used for treatment of trauma, disease or injury of bone tissue.⁶ 

For any utilization of a material we must understand its features and properties. All glasses, whether manufactured via melt quenching, sol–gel processing or some other suitable method, possess two common characteristics: the glass structure and the gradual change of several properties when heated or cooled between the solid form, glass, and the liquid form, melt. Glass is often defined as “an amorphous solid completely lacking long range, periodic arrangement of the atomic structure, and exhibiting a region of glass transformation behaviour”.⁷  In this transformation the properties of the liquid gradually change into solid state properties. Some general rules of glass formation and the high temperature properties to be considered in glass forming processes are explained below. A detailed discussion of glass structure and its influence on glass properties is given in Chapter 3.

Melt capable of forming a glass maintains its liquid-like structure as a supercooled liquid below the melting point of the crystal, to transform into a brittle, elastic glass on further cooling. The transformation behaviour of a glass forming melt is depictured in Figure 1.1. The figure gives changes in the volume or the enthalpy at characteristic temperatures when the melt converts into either a crystalline state or forms a supercooled melt which transforms into a non-crystalline solid, a glass, on continued cooling. Glass transformation takes place at a certain temperature called the glass transition temperature or glass transformation temperature T_g. This temperature depends on the composition of the glass but also on the cooling rate; the higher the rate the higher the transformation temperature. Glass transformation is thus a time-dependent behaviour and the value of T_g is always dependent on the thermal history of the glass.

Glass transformation temperature is measured normally from graphs recorded in thermal analysis or dilatometry of glasses, i.e. in heating glasses from the solid state into the molten state. The T_g value obtained depends on the heating rate, the instrumental method used in the measurement, and the thermal history of the glass. Below T_g the glass behaves as an elastic solid, around T_g the glass shows viscoelastic behaviour, and above T_g the glass softens and starts to behave as a viscous liquid. The glass transformation temperature is thus an important parameter for selecting the experimental conditions in thermal treatment of a particular composition into various products above T_g. In addition, the annealing curve, i.e. controlled time–temperature history for relaxing any stresses present in the glass after the forming operations, is based on the T_g value. For bioactive glasses, the T_g value is of interest when sintering porous implants from powdered fractions or when estimating a suitable annealing curve for a monolith. However, T_g has no practical significance, e.g. in the quenching of glass melts into water to give glass particles. In contrast, if internal stresses in the particles caused by the quenching are to be relaxed, then an additional annealing step at around T_g is required.

If the melt were to crystallize during the cooling, abrupt change in the properties takes place when the crystal, with long range, periodic arrangements of the atoms, forms at the melting point (T_m in Figure 1.1), independently of the cooling rate. Since glass forming liquids have typically high viscosity at the melting point they easily form a supercooled liquid on cooling. For glass forming liquids, the melting point is often referred to as the liquidus temperature T_liq. The liquidus temperature is the highest temperature at which crystals can be in thermodynamic equilibrium with the melt. Thus, the value of T_liq (or T_m) does not depend on the melting history of the glass. From the manufacturing point of view, there is always a risk of crystallization if the melt is processed below T_liq. In commercial soda lime–silica melts, crystals form within a few hours if held below the liquidus temperature.⁸ 

The liquidus temperature is traditionally measured from glass samples treated in a furnace with controlled temperature profile for periods long enough to precipitate crystals which can be identified optically in quenched samples. The high tendency of bioactive glasses to crystallize in thermal treatments combined with the slow melting or dissolution of the crystals eventually formed during the heating questions the accurate determination of liquidus temperature in bioactive glasses. Thermal analysis is frequently used to estimate the temperature range in which the crystals melt.

A defect-free amorphous glass contains no crystals. Nevertheless, all glasses crystallize but at different rates in the temperature window between glass transition T_g and liquidus T_liq. Detailed understanding of crystallization is vital in fabrication processes requiring any longer thermal treatment in this critical temperature window. Although bioactive silicate glasses and soda lime–silica glasses are composed to a large extent from the same oxides, the markedly lower content of the main glass network forming oxide, SiO₂, in bioactive silicate glasses gives easy crystallization.

Crystallization of a liquid happens via two processes: nucleation and crystal growth. In nucleation, a sufficient quantity of atoms form an ordered first structure (the nucleus) after which crystal growth is facilitated by new layers of atoms forming around the nucleus. For crystalline solids, these processes take place at the melting point, while for typical glass forming melts both nucleation and crystal growth show maximum values within certain temperature ranges below the melting point (Figure 1.2). All melt-derived glasses pass this critical crystallization range during the fabrication. The overlapping of the rate curves depends on glass composition and partly controls the suitability of a particular composition to thermal treatments in the temperature range from glass transition to melting point.

Detailed information on crystallization mechanism, i.e. nucleation and crystal growth parameters, is available only for a few melt-derived bioactive glass compositions. In this chapter, the results from measurements of the characteristic values for bioactive glasses 45S5, S53P4 and 13-93 are discussed. Table 1.1 gives the oxide compositions of these three glasses. The original bioactive glass 45S5 Bioglass^® developed by Professor Hench et al.²  and S53P4 BonAlive^® developed by Andersson et al.⁹  are commercially available for certain clinical applications. The clinical applications of bioactive glasses are discussed in Chapters 14 and 19. Glass 13-93 was originally developed by Brink et al.^10,11  to enable easier fabrication into shapes that are challenging from the glass manufacturing point of view. Since then, glass 13-93 has been used to manufacture continuous fibres and porous sintered implants or scaffolds by several research groups.^12–21 

Crystallization of S53P4 and 13-93 takes place via surface nucleation and crystal growth mechanisms while for 45S5 the crystallization is more complex.^22–24  For 45S5, smaller and larger particles are reported to show different nucleation mechanisms,^23–25  and the crystallization proceeds rapidly from surface to the bulk.^23,26  Phase separation above T_g into immiscible liquids precedes the nucleation in 45S5.^25–27 

The crystallization mechanism is most often based on activation energy values, Johnson–Mehl–Avrami values, and nucleation-like curves determined from thermal spectra which are measured using certain temperature–time treatment in thermal analysis.^28–30  The maximum nucleation rate for the primary crystals which form above T_g was around 560–580 °C for 45S5, while a slightly higher temperature of around 610 °C was suggested for S53P4.²³  Correspondingly, the highest nucleation rate took place at 700 °C for 13-93.²²  Crystal growth rate is markedly lower in 13-93 than in the two other compositions. In thermal treatment of 45S5 and S53P4 more than one crystalline phase may form depending on the temperature.^23–27  Since this chapter focuses on properties and characterization of bioactive glasses, the crystallization is given only as a limiting factor for the fabrication window, while the phase separation and formation on secondary crystals are not considered.

The overall crystallization procedure is usually determined using differential thermal analysis or differential scanning calorimetry. A certain size fraction of glass particles is heated using a controlled rate (5–20 °C min⁻¹) and thermal effects compared to an inert reference material are recorded as a function of temperature. Figure 1.3 shows typical thermal spectra for the three bioactive glasses 45S5, S53P4 and 13-93. The endothermic peak in the thermal spectra at around 500–600 °C gives the glass transformation while the exothermic peaks at around 600–800 °C (45S5 and S53P4) and 800–1000 °C (13-93) give the temperature spans at which crystals form upon heating.

The last thermal effects in Figure 1.3 are associated with the melting of the crystal, i.e. T_liq of the glasses. The thermal effect related to the liquidus is typically low and thus not always clearly shown in the thermal spectrum. The values for the thermal effects in Figure 1.3 and values reported in the literature for the three glasses are summarised in Table 1.2. The minor differences in the measured values can be explained by differences in the experimental conditions (thermal history of the glass, heating rate, particle size) and using the inflection point, the first deviation from baseline or the maximum value of the thermal effect to describe the property value.

Glass stability is an often-used indicator describing the resistance to crystallization during heating. Glass stability is typically given as the difference between the crystallization onset value (T_x) and the glass transition temperature (T_g). Figure 1.3 and Table 1.3 give the stability ranges of 88 °C for 45S5, 130 °C for S53P4, and 246 °C for 13-93. Accordingly, 13-93 can be thermally treated within a wide temperature range without crystallization while 45S5, showing a temperature difference less than 100 °C, crystallises easily. Correspondingly, S53P4 allows limited sintering above T_g.³⁴ 

Ultimately, crystallization during heating is a kinetic phenomenon and suitable processing parameters can be given as time–temperature–transformation (TTT) diagrams giving the maximum allowed duration at various temperatures to avoid crystallization. The TTT curves of 45S5 indicate an increase of transformation from 10% crystals after 200 s to 90% after 2000 s at 622 °C.²⁶  The transformation rate increases with temperature: 90% of crystals were measured already after 250 s just at around 680 °C.²⁶  No detailed TTT curves have been published for S53P4. The crystallization studies indicate that S53P4 has a markedly lower crystallization rate than 45S5. No crystals were detected in samples sintered of S53P4 at 650 °C for one hour while at 700 °C a distinct, crystallized layer had formed at the particle surfaces.³⁴  The crystallization parameters for 13-93 have been characterized in more detail.²²  The crystal growth was very slow still at 800 °C but rapidly increased above 850 °C.

The primary crystals, which are formed during heating of bioactive glasses belong mainly to two different composition areas: crystals consisting of Na₂O, CaO and SiO₂ (NCS) in various ratios, and crystals consisting of CaO and SiO₂ (CS).³³  In addition, secondary crystals may form at higher temperatures.^{23–27,34,37}  Though the exact crystal composition is of relevance in the manufacture of glass-ceramics, the primary crystal type, NCS or CS, correlates with the crystallization tendency of the glass.³³  The glasses forming NCS type crystals are much more sensitive to thermal treatments than the glasses forming CS type crystals. X-ray measurements on thermally treated glasses are used to characterize the exact phase composition of the crystals.

The basis of glass forming processes is the flow of the melt and heat transfer to complete the forming without deformation under gravity after the required shape is reached. The viscosity of the melt and its change with temperature are used to estimate the suitability of a particular melt composition to various forming operations. The viscosity of glass melts varies strongly with temperature but is also influenced by the glass composition. Since bioactive glass melts achieve optimal viscosity values for different forming operations at temperatures close to the crystallization range, time dependence must also be considered in the manufacture of crystal-free products. The standard viscosity points for soda lime–silica glasses are often used to describe the viscosity–temperature behaviour. For the manufacture of bioactive glasses, the following viscosity standard points are of interest: glass melting at 10^2.0 dPa s, glass working temperature at 10⁴ dPa s, glass transformation temperature around 10^11.3 dPa s, and annealing around 10¹³ dPa s.⁸  Several additional standard points are defined for soda lime–silica glasses between the glass transformation and liquidus temperatures.⁸  However, these viscosity standard points are irrelevant for bioactive glasses with a strong crystallization tendency.

For bioactive glass products based on melt-derived compositions, two forming routes are obvious: (i) direct forming from a viscous melt into the desired product either via casting into moulds, quenching into water, or drawing into continuous fibres, and (ii) reheating glass above T_g to allow sintering of particles together into desired porous architectures, drawing of fibres from a pre-form or marbles, or sealing particles to give a layer on a substrate. In general, crystallization below the liquidus is slow for melts with high viscosity at T_liq. Thus, high viscosity at T_liq favours forming processes.

Today, most commercial melt-derived bioactive glasses are manufactured as particles of different size fractions, or as rather small monoliths. Powdered fractions are often obtained via quenching the melt into water to give small granules, which are then dried and sieved to desired size fractions. If manufactured via crushing of monoliths, rapid transfer into the annealing furnace is required after casting to avoid crystallization. However, there are no strict viscosity requirements for the glass melts used to produce particles. In contrast, in casting monolithic implants, a rapid increase of the viscosity guarantees geometric stability at the high temperatures immediately following the forming.

Drawing of continuous fibres from bioactive glass melts is challenging. Fibres would be of interest as components in composites or other special structures utilising their very high surface area to induce bioactivity.

Drawing of continuous fibres from commercial fibreglass is characterized by a rather narrow viscosity range of 10^3.5–10^4.0 dPa s.⁸  The difference between the fibre-forming and liquidus temperatures should be at least 50 °C to avoid crystallization during the drawing. The forming can be performed below the liquidus if the nucleation and crystal growth temperatures are not overlapping to a great extent. If the nucleation temperature range is well below the crystal growth temperature range, fibre drawing can be realized below the liquidus.

Fibres are also drawn from a pre-form or from marbles, i.e. glass reheated to the fibre drawing temperature corresponding to the optimal viscosity range of 10^3.5–10⁴ dPa s.⁸  During the heating, the glass passes through the nucleation temperature range. If nuclei are easily formed, fibres can be drawn only over a limited period of time. The nuclei and growing crystals may also affect the composition of the fibres manufactured from a pre-form of a given composition. The strong crystallization tendency between T_x and T_liq limits the suitability of bioactive glasses to the fibre-drawing processes.

Viscosity also controls the suitability of a certain glass composition to sintering of amorphous porous scaffolds. A porous tissue engineering scaffold has an open continuous porosity which supports ingrowth of fully vascularised new tissue. A suitable viscosity range in sintering of glass particles into porous bodies via viscous flow is 10⁸–10^8.8 dPa s.⁸  When manufacturing the scaffolds via the sintering route, the particle size is one critical factor for the commencement of the densification process. Careful optimisation of the time–temperature history is required for achieving defect-free scaffolds with adequate strength.

It is well known that 45S5 and S53P4 are prone to crystallization during any prolonged thermal processing in the temperature window between the glass transition and the liquidus temperatures. Accordingly, the commercial products of these glasses are typically melt-quenched products or small monoliths which do not require any specific processing within the critical temperature window.⁶  In contrast, glass 13-93 is one composition in a series of 30 glasses that were designed to enable hot-working.¹⁰  In vivo studies confirmed that this composition is not only suitable for versatile thermal treatments but also shows similar reactivity with bone as glass S53P4.^9,11,39  Later, this composition was frequently studied in various tissue engineering applications as a composition allowing for manufacture of amorphous products without crystallization.^12–21  Viscosity measurements explain the differences in the suitability of the three glasses 45S5, S53P4 and 13-93 to various hot-working methods. Since the viscosities vary by several orders of magnitude between the practical melting temperatures (10^2.0 dPa s at 1300–1400 °C) and T_g (10^11.3 dPa s at 500–600 °C), several methods are needed to describe the viscosity at different temperature ranges.

The low temperature viscosities of bioactive glasses have been measured for increasing temperatures above T_g using beam bending viscometry⁴⁰  and hot-stage microscopy.^10,41  The measured viscosity–temperature values for 45S5, S53P4 and 13-93 are shown in Figure 1.4. Only a few values could be measured for 45S5 and S53P4 using hot-stage microscopy while for 13-93 several values were obtained.⁴¹  For each composition, crystallization interrupted the measurements at increasing temperature. The high temperature viscosities were measured using a rotation viscometer for decreasing temperatures (Figure 1.4).^40,41  Also at the high temperature range, the viscosity of 13-93 could be measured for a clearly higher range than for 45S5 and S53P4. Between the low and high temperature ranges, the glasses exist as mixtures of varying amounts of melt and crystals and no true viscosity for the melt can be specified. The ranges for sintering and fibre drawing are also marked in Figure 1.4. 45S5 melt crystallizes at very low viscosity during cooling. Accordingly, normal fibre drawing processes cannot be used to achieve continuous fibres of 45S5. In contrast, the 13-93 melt is resistant to crystallization below the liquidus down to around typical fibre drawing viscosities. Similarly, 13-93 can be sintered into amorphous porous bodies, while 45S5 and S53P4 easily crystallize.

Today, bioactive glasses are mainly used to fill defects and to promote and support bone tissue regeneration. The current research trend is to fabricate glass and glass-ceramic scaffolds with architectures mimicking the three-dimensional interconnected porosity of natural bone.^42–49  The porous scaffold not only actively induces bone regeneration but, optimally, also degrades at a rate matching the tissue ingrowth rate. The requirements of the porous tissue engineering scaffolds are discussed in detail in Chapter 21. In general, adequate mechanical strength and controlled, predictable reaction and dissolution rates are the most important properties for the bioactive glass product in its final application.

The theoretical strength of flawless solid silicate glass is ≈35 GPa.^50,51  However, the ability of a glass to resist fracture when a crack is present, i.e. the fracture toughness, is low. Thus, already small flaws decrease the strength noticeably and typical strengths of common glass products are only around 14–70 MPa.⁵¹  Glasses are brittle and fail without yielding as indicated by the high Young's modulus of silicate glasses, 45–100 GPa.⁵²  Table 1.3 summarizes typical mechanical property values reported for glass 45S5 and bone.^{42,46,51,53–55} 

Mechanical durability is a major drawback for porous bioactive glass scaffolds both because of the intrinsic brittleness of glasses and the high interconnected porosity required for cell penetration, vascularization, and nutrient flow.^43–49  Thin struts, large pore size, and high total porosity may give scaffolds which have too low mechanical strength even in low-load applications. The strength must also be high enough to sustain the surgery without mechanical failure. After implantation, the scaffold starts to dissolve which may lead to crack formation in the struts especially when subjected to a load or mechanical stresses.

Since most melt-derived glasses partly crystallize in the sintering processes applied in scaffold fabrication, most data available on mechanical properties are for glass-ceramics. Mechanical properties of scaffolds prepared via different approaches, such as the foam replication technique and various 3D printing techniques from the parent glass 45S5, have been reported in several studies.^{42–44,55–60}  In contrast, only limited information is available on porous scaffolds sintered of glass S53P4.³⁴  The mechanical properties of amorphous scaffolds based on glass 13-93 have been tested frequently in vitro and also in vivo.^{16,21,44–46,61–66}  The porous scaffolds are often coated with a biodegradable polymer for increased mechanical strength. The degradation of mechanical strength in vivo depends on several variables: the composition of the glass, the pore architecture and total porosity of the scaffold, the degradation rate of the polymer coating, and the effects of the interaction of the polymer and glass on the degradation of both materials. One additional paramount factor determining the suitability of the scaffold to the regeneration of loaded bone is the influence of the rate of new bone formation on the overall mechanical properties of the implant. The strength properties of the porous scaffolds are discussed in Chapter 21.

Bioactive glasses are designed to dissolve and react in a controlled manner in the body environment. Traditionally, bioactive glasses have been described as surface-active materials capable of forming a mechanically strong chemical bond with living tissue, mainly bone.^4,6,45,67  The basis for the glass to form chemical bonding with tissue is the time-dependent dissolution and precipitation reactions of the glass with its surrounding solution. The changes in the ion concentrations of the solution around the reacting glass lead to the formation of the carbonated hydroxyapatite (HCA) interfacial layer. The HCA layer is biomimetic with the inorganic mineral apatite in the bone and gives thus a chemical fixation of the implant with the surrounding tissue instead of a mechanical fixation provided by the fibrous capsules around biostable implants. Biomimetic crystals not only have a similar composition to the apatite in living bone but also their crystallite size is compatible with the nano-sized bone or dentine apatite crystals.^68,69  The principles of the bone bonding mechanisms have been described in several papers.^2,45,67,70 

The HCA layer surface also makes the glass osteoconductive. Since the bioactive glasses dissolve with time while supporting apatite precipitation at the dissolving surface and nearby tissue, the bioactive glasses can also be classified as resorbable materials. The ion release products from the bioactive glasses are known to activate several families of genes, among others the genes that regulate osteogenesis and the production of growth factors, to affect adsorption of proteins and cell attachment at the surface of the reacting glass.^71–75  Thus, bioactive glasses, which stimulate regeneration of new bone via ionic dissolution products, are not only osteoconductive but also osteostimulative. The dissolution and precipitation reactions of bioactive glasses in vitro and in vivo lead to similar surface template structures, which favor precipitation of hydroxyapatite.^9,76  Accordingly, in vitro formation of HCA at the glass surface is usually taken as an indication of the bioactivity.⁷⁷  The glass surface first partly dissolves and then serves as a substrate for precipitation of HCA. Primarily, the precipitate is amorphous but converts into carbonated hydroxyapatite with time.

Interestingly, the utilization of bioactive glasses is not restricted only to applications where bone bonding provided by HCA crystals is desired, i.e. bone tissue engineering. The impact of the released ions on activation stimulation of genes and cells involved in hard and soft tissue regeneration, wound healing, and angiogenesis has been discussed in several papers.^{4,6,45,78–80}  Bioactive glasses have potential also in the regeneration of cardiac tissue, lung tissue, nerves, gastrointestinal tissue, etc.⁵  The HCA layer which forms at the surface of the bioactive glass provides a bonding interface also to soft tissues. However, the mechanisms for interactions between the bioactive glass and the cells that compose the different soft tissues are still poorly understood.

The pH and the ion concentrations of the interfacial solution increase rapidly around the dissolving glass especially when using high concentrations and/or small particles. This pH effect and the released ions inhibit and prevent bacterial growth around the glass.^81–91  The antibacterial effect may be enhanced by doping the glass with Ag, Ce, Co, Zn, etc.⁷⁹  These findings indicate that a detailed knowledge of the dissolution mechanisms is also essential when assessing the antibacterial effects of various size fractions of bioactive glasses.

The glass dissolution reaction mechanisms are well described in the literature.^3,92–101  Glass can either dissolve congruently, i.e. uniformly, which implies that the ratios of elements in the solution are the same as in the dissolving glass, or incongruently, i.e. preferential or selective leaching of some of the elements. The reaction mechanisms depend on the glass composition and environmental conditions, such as the surface area to volume ratio, and on the pH of the solution. In general, all silicate glasses dissolve in aqueous solutions but the extent of the reactions strongly depends on the composition of the glass and solution. In more general terms, these basic dissolution mechanisms can be divided into primary reactions, leading to the formation of a subsurface zone consisting of a silica-rich layer, and secondary precipitation reactions, giving an outer surface layer or an alternate surface layer structure. These reaction mechanisms also explain the formation of the typical dual layer at the bioactive glasses.

The reactions leading to the formation of the bone bonding are usually described as a sequence of eleven reaction stages at the surface of the glass.⁹⁷  The first five rapid reaction stages take place at the surface of a bioactive glass with the highest level of bioactivity within 24 h. These reactions explain the formation of a dual layer of hydrated silica (silica-rich gel) and polycrystalline HCA at the glass surface. At the subsequent reaction stages, bonding with bone via a series of biological reactions takes place and within 6 to 12 days the final product, a collagen-HCA matrix containing mature osteocytes, is formed.⁹⁷  Simplified, the HCA at the glass surface is similar to the apatite in natural bone and thus is able to form mechanically strong chemical bonding with the living hard tissue. The five first bioactivity reactions stages are summarized below and in Figure 1.5.

At pH <9, a selective leaching characterized by an ion exchange reaction of alkali and alkaline-earth ions with protons from the surrounding solution dominates (Reaction I and Figure 1.5, insert 1). The field strength and the radius of the network-modifying mobile ions affect their leaching rate.³  Accordingly, alkali ions and the more weakly bonded ions with larger radii are leached out more easily than the smaller ions. The silica network structure may also react with water via hydration (Reaction 2 and Figure 1.5, insert 2). The ion exchange and the hydration reactions result in a slightly porous hydrated silica-rich layer and an increase in the pH of the surrounding solution. The hydroxyl ions act as a catalyst for network dissolution and, at pH>9, the hydrolysis reaction starts to dominate leading to depolymerisation of the glass network structure (Reaction 3, Figure 1.5, insert 3). The silanol groups (Si–OH) may gradually depolymerise further and release soluble silica to the solution. The silanol groups can also condense to give a silica-rich gel at the glass surface (Reaction 4, Figure 1.5, insert 3).

The sub-surface structure that develops in the aforementioned reactions is often described as a silica-rich gel layer as it may be porous and contains water and silanol groups. The secondary reaction type, precipitation, starts when the solubility of the dissolved ions is exceeded. The silica–rich gel provides suitable nucleation sites for Ca²⁺ and PO₄³⁻ migrating from the bulk glass or precipitating from the solution to form the amorphous calcium phosphate film, which gradually crystallizes into HCA (Figure 1.5. insert 4). Also, other anions, such as F⁻, OH⁻, and CO₃²⁻, may be incorporated into the crystalline structure.⁹⁷  This crystallized HCA is similar to that in natural bone and it is able to form chemical bonds with natural bone via a series of biological reactions. The formation of an HCA layer on the leached silica-rich layer may act as a diffusion barrier, depending on the density of the layer. This decreases the driving force for further dissolution and might partially help explain why, with some bioactive glass compositions, intact core glass is still found after several years of implantation.¹⁰³ 

For melt-derived glasses, the layered surface structure of SiO₂ and HCA forms rapidly on glasses with low silica content. The network structure of these glasses consists mainly of Q² and Q³ units, i.e. an open silicate network with poor chemical durability.^104,105  However, composition alone does not determine the bioactivity, the sample form, particle size, concentration of the particles in the solution, fluid flow, etc. also affect the bioactivity expressed as HCA formation in vitro or bone bonding in vivo.

Several analysis methods have been utilized to gain information on the dissolution behaviour and reactions of bioactive glasses in vitro. These are based on studying the changes occurring both in the glass and in the solution upon incubation of the glass samples in phosphate buffered aqueous solutions, such as a simulated body fluid.

Recently, the Technical Committee 4 (TC04) of the International Commission of Glass (ICG) suggested a static in vitro method for testing the apatite-forming ability of bioactive glasses.¹⁰⁶  In the test, glass powders are immersed in simulated body fluid (SBF) using a fixed mass to volume solution ratio in airtight polyethylene containers. The containers are held at 37 °C in an incubating orbital shaker for various time periods. At the end of each immersion time period, the particles and solution are separated and analysed. The ion concentrations in the solution are measured using an inductively coupled plasma optical emission spectrometer (ICP-OES). In addition, the pH of the solution is measured. The glass samples are characterized using Fourier transform infrared spectrometry (FTIR), X-ray diffraction (XRD), and/or scanning electron microscopy (SEM). The method was tested at several laboratories and was found to give appropriate values for apatite formation, especially on samples consisting of particles or glasses with high surface area. Time points for detection of HCA nucleation on the sample surface and changes in phosphate concentration were found to correlate. The test method is recommended when comparing new glasses with known compositions reported in the literature.¹⁰⁶ 

Dissolution tests in static or agitated solutions give comparable values between different glass compositions but do not provide accurate information of the glass dissolution kinetics. However, the human body is not a static environment and the fluid flow conditions vary in different locations. Different continuous in vitro testing methods have been developed to better imitate the conditions in the dynamic body environment. These have utilized fluid flow rates ranging from 0.03 mL min⁻¹ to 2 mL min⁻¹ to study the in vitro reactions.^39,107–113  In most studies, the solution flow slowly changes the fluid in a relatively large container containing the sample. In several studies, the same solution has been circulated in the system causing changes to the original solution composition, but fresh solutions have also been used. The solution has been analysed for samples collected at predetermined time intervals. By connecting a dynamic measurement method directly into a fast analysis method, such as pH or ion measurement, on-line data can be collected. Inductively coupled plasma optical emission spectroscopy (ICP-OES) offers a sensitive on-line method for qualitative and quantitative determination of metals and certain non-metals in solution.^39,113 

Three phenomena are of interest when assessing the in vitro bioactivity of glasses: (i) the increase in the pH of the fluid around the glass, (ii) the ion concentrations in the fluid of the ions releasing from the dissolving glass, and (iii) the layers that form at the dissolving glass surface.

The layer formation on 45S5, S53P4 and 13-93 has been reported frequently but seldom in the same studies or using the same experimental conditions. Although typical HCA and Si-rich surface layers form on each composition, clearly thicker layers form rapidly on 45S5, while the layer thickness is least on 13-93.^39,114–116 

The ability of a glass to form reaction layers correlates with the release of ions from the glass and changes in the pH of the immersion solution.³⁹  Figure 1.6 shows the initial release of Ca and Si into a fresh TRIS buffer solution fed continuously through a bed of particles of 45S5, S53P4 and 13-93.³⁹  The solution flow rate of 0.2 mL min⁻¹ is close to the value reported for fluid flow in human muscles.¹⁰⁸ 

The amount of Na ions released from 45S5 was very high and the values exceeded the detection limit. For S53P4, the Na release values were also close to the limit but decreased after the initial peak value to around 110 mg L⁻¹. Lower amounts of Na dissolved from 13-93, around 10 mg L⁻¹, while the amount of K was 20 mg L⁻¹ at 15 min. For all glasses, the concentration of Ca ions stabilized after a minor initial peak and was around 130 mg L⁻¹ for 45S5, 100 mg L⁻¹ for S53P4, and 50 mg L⁻¹ for 13-93 at 15 min. At this time point, the concentration of Mg ions releasing from 13-93 was around 9 mg L⁻¹. The initial release of soluble Si ions was highest from 45S5 while the concentration increased steadily for S53P4 and 13-93. Finally, the initial phosphate release was around 10 mg L⁻¹ for 45S5 and S53P4 while slightly lower concentrations were measured for 13-93.

The pH values were recorded in similar experimental conditions using a flow-through micro volume pH electrode. The pH curves for the glasses verify the highest overall ion dissolution from 45S5 and would also be typical for systems in which reaction layers have been observed (Figure 1.7).

The initial ion dissolution and pH profiles were in line with the observations of the in vitro and in vivo layer formation on these glasses. The high initial release of calcium ions from 45S5 suggests that the saturation limit in the solution is exceeded already after the first minutes of the contact with the 45S5 particles and leads to rapid calcium phosphate precipitation while slower precipitation is likely on the other glasses. The reaction rate also depends on the environmental conditions such as the surface area to volume ratio or flow rate of the solution, both of which affect the pH of the solution.

The initial ion release can also be used to estimate the capability of the glasses to affect cellular processes. For osteostimulation, the release rates of biologically active soluble Ca and Si ions are 60–90 ppm for Ca and 15–30 ppm for Si.^71,72,78  If present in a critical concentration, these ions also activate, or up-regulate, seven families of genes in osteogenic cells.^71–73,78  The initial concentrations released from 45S5 and S53P4 into the TRIS buffer solution were higher than these critical values while the calcium concentration was less than the critical concentration (Figure 1.5). Whether the ion concentrations stay at critical levels for long enough time periods should be verified with longer test times. Relatively high concentrations of potassium ions were released from 13-93 during the first minutes. Extensive potassium release may limit the utilization of 13-93 or other potassium oxide containing glasses in product forms giving high surface area to volume ratio to avoid any deleterious cellular effects.¹¹⁷  The dynamic dissolution tests give additional interesting information for assessing the dissolution kinetics and measuring the ion release rates from bioactive glasses.

Melt-derived bioactive glasses are commercially utilized mainly as granules or particles to repair and treat bone injuries and defects. In contrast, extensive research effort is devoted to developing porous scaffolds for soft and bone tissue regeneration. Thorough understanding of the material glass, its physical and chemical properties is essential when designing novel scaffolds and devices with desired porosity, shape, and controlled, pre-determined ion release. Fabrication of monoliths, porous scaffolds or fibres is limited by the viscosity–temperature behaviour and crystallization tendency of the glass in hot-working. Biological responses of the glass can be estimated by studying the ability of the glass to form layered structures of silica and hydroxyapatite at the surface in vitro. Ion release rates from the glass give additional information for estimating the dissolution rate of the glass and its capability to stimulate and support tissue regeneration.