It can be easily noticed that experimental proof supporting one or the other
concepts of the internal glass structure is mainly based on X-ray diffraction
methods. However, taking into account that standard diffraction methods yield
results on average atom distribution, the application of these methods for studying
glass structures is very limited [11]. On the basis of these methods, only the lack of
long-range order can be definitively determined in the studied glasses; however,
determining homogeneity (Zachariasen–Warren theory) or heterogeneity (Lebedev
theory) inside glass is very difficult or even impossible. The development of other
diffraction techniques, e.g., with the utilization of neutron beams, X-ray fluorescence analysis, etc., would enable the determination of several glass network
parameters. Mozzi and Warren [12] determined on the basis of X-ray fluorescence
analysis that the length of the Si–O and O–O distance in m-SiO 2 are 1.62 and
2.65 Å, respectively. Si–O–Si bond angles exhibit a normal statistical distribution
with a maximum around 152°. However, there is less variety in the values of O–Si–
O angles around the tetrahedron angle equal to 109.5°. These values are comparable
to those determined for several crystalline silicates [13, 14]. The initial advantage of
the continuous uninterrupted network model was due to its support from more
precise and more graphic experiments [8–10]. On the other hand, many authors
[15–17] have called attention to the presence of characteristic points in property–
chemical composition or property–property curves, at which these curves change
their course. These points are usually associated with the boundaries of a certain
phase existing, or as in the case of the refractive index–temperature curve, the
temperature of cristobalite polymorphic changes. These phenomena, much like
many others, such as the presence of voids in glass (greater permeability of gases in
glass than in the respective crystal), cannot be explained by the Zachariasen–
Warren theory. All these phenomena can be explained under the assumption that
regions with chemical composition and structure similar to respective crystalline
substances are present in the glass structure. The existence of such regions (domains) has been unambiguously confirmed by means of vibrational spectroscopy
studies (e.g., [18–22]), which have demonstrated a great similarity between m-SiO 2
MIR spectra and those obtained for high-temperature cristobalite. Therefore, current
views on glass structure lean more in the direction of the crystallite theory; however, the predominant opinion is that the structure of real glasses is somewhere in
between, with characteristics corresponding to one of the theories [16]. The range
of order is currently the subject of the most controversy; however, most authors
agree that short-range order is close to respective crystalline structures. Assuming
that middle-range order is present in silicate glass, there remains the question of the
type of order within domains and whether or not the domains only have one
characteristic (homogeneous build) or different kinds of order (heterogeneous build)
(Fig. 15.2).
In the literature pertaining to m-SiO 2 structure, many pseudo-crystalline models
can be found postulating the existence of ordered regions (domains) with different
SiO 2 polymorphs (e.g., [11, 23–29]). These are most common models that assume
homogeneous build, i.e., with domains exhibiting only one characteristic order.
Still, there is no universal consent among scientists pertaining to the characteristics
460
W. Mozgawa et al.
concepts of the internal glass structure is mainly based on X-ray diffraction
methods. However, taking into account that standard diffraction methods yield
results on average atom distribution, the application of these methods for studying
glass structures is very limited [11]. On the basis of these methods, only the lack of
long-range order can be definitively determined in the studied glasses; however,
determining homogeneity (Zachariasen–Warren theory) or heterogeneity (Lebedev
theory) inside glass is very difficult or even impossible. The development of other
diffraction techniques, e.g., with the utilization of neutron beams, X-ray fluorescence analysis, etc., would enable the determination of several glass network
parameters. Mozzi and Warren [12] determined on the basis of X-ray fluorescence
analysis that the length of the Si–O and O–O distance in m-SiO 2 are 1.62 and
2.65 Å, respectively. Si–O–Si bond angles exhibit a normal statistical distribution
with a maximum around 152°. However, there is less variety in the values of O–Si–
O angles around the tetrahedron angle equal to 109.5°. These values are comparable
to those determined for several crystalline silicates [13, 14]. The initial advantage of
the continuous uninterrupted network model was due to its support from more
precise and more graphic experiments [8–10]. On the other hand, many authors
[15–17] have called attention to the presence of characteristic points in property–
chemical composition or property–property curves, at which these curves change
their course. These points are usually associated with the boundaries of a certain
phase existing, or as in the case of the refractive index–temperature curve, the
temperature of cristobalite polymorphic changes. These phenomena, much like
many others, such as the presence of voids in glass (greater permeability of gases in
glass than in the respective crystal), cannot be explained by the Zachariasen–
Warren theory. All these phenomena can be explained under the assumption that
regions with chemical composition and structure similar to respective crystalline
substances are present in the glass structure. The existence of such regions (domains) has been unambiguously confirmed by means of vibrational spectroscopy
studies (e.g., [18–22]), which have demonstrated a great similarity between m-SiO 2
MIR spectra and those obtained for high-temperature cristobalite. Therefore, current
views on glass structure lean more in the direction of the crystallite theory; however, the predominant opinion is that the structure of real glasses is somewhere in
between, with characteristics corresponding to one of the theories [16]. The range
of order is currently the subject of the most controversy; however, most authors
agree that short-range order is close to respective crystalline structures. Assuming
that middle-range order is present in silicate glass, there remains the question of the
type of order within domains and whether or not the domains only have one
characteristic (homogeneous build) or different kinds of order (heterogeneous build)
(Fig. 15.2).
In the literature pertaining to m-SiO 2 structure, many pseudo-crystalline models
can be found postulating the existence of ordered regions (domains) with different
SiO 2 polymorphs (e.g., [11, 23–29]). These are most common models that assume
homogeneous build, i.e., with domains exhibiting only one characteristic order.
Still, there is no universal consent among scientists pertaining to the characteristics
460
W. Mozgawa et al.