LECTURE 2 :

STRUCTURE AND EXPERIMENTAL

CHARACTERIZATION
- Order in glasses
- Zachariasen random network theory
- Glass families and glass forming ability
- Experimental tools for structural characterization

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 A) SHORT, INTERMEDIATE AND LONG RANGE ORDER IN GLASSES
Bragg's Law: conditions for constructive interference, some angle of incident x-rays that
will produce intense diffraction peaks associated with the crystalline lattice spacings.
2
This situation is not met in amorphous systems (glasses) because of the lack of
periodic reproduction of an elementary cell. Broadening of diffraction peaks.

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

V. Simonet, R. Bellissent, J. Physique (2003)
 Pair distribution function g(r) # probability to find another atom at a distance r
from a central atom

1rst coordination
shell
2d coordination
shell

dist. between successive neighbours

Can something be said beyond the first peak of g(r) ?

Short range order : first shell of neighbors (coordination number)

Lengthscale ~ atomic bond distance (2-5 A)

Intermediate range order : the way elementary building blocks arrange together
[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE
 Intermediate range order : the way elementary building blocks arrange together

The basic building block of SiO2 is the

SiO4/2 tetrahedron (identical to -quartz)
a2
2 Disorder is due to the way the tetrahedra
connect together.
Angle (Si-O-Si) and 2 (dihedral angle).
Elements of intermediate range order (<10 A)

Long-range order : beyond 10 A Ge15Te85

Ring statistics
Computer simulations
Experiments (Raman, rare systems) Kalikka et al, JPCM 2011

Glasses have random networks with the short range order of corresponding crystals

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 B) ARE GLASSES CONTINUOUS RANDOM NETWORKS ?

Zachariasen's Random Network Theory ( 1932)

Before Zachariasen, the structure of a glass was believed to be comprised of

nanocrystals (~20 size) given the broadening of the diffraction peaks.

Scherrer equation : monocrystal k=0, polycrystal k finite and the size

of the grains is L# 1/k.

He noted similar mechanical properties (elastic modulus, etc.) between glasses

and crystals and so expected similar structural energies

Glasses lack the periodic (long range) order of a crystal. Infinite unit cell (no
repeating large scale structures). 3D network lacking symmetry and periodicity

ISOTROPIC: same average packing and properties in all directions. Different from
crystals (directions).

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 Zachariasen's Rules ( 1932) for Glass formation based on an empirical observation on
oxides.

1. No oxygen atom may be linked to more than two cations

2. The cation coordination number CN is small: 3 or 4.
3. Oxygen polyhedra share corners, not edges or faces.
4. For 3D networks, at least three corners must be shared

In general, all four rules should be satisfied for glass formation to occur.
Low coordination numbers, corner-sharing rules imply that glass formation is more likely
with open, low density polyhedral structures.

Having stated this, one identifies a certain number of compounds to be either good glass
formers (i.e. satisfying the 4 rules) or bad glass formers (do not satisfying at least one of
the rules).

Example : SiO2, GeO2 are good glass formers (CN =4)

V2O5 (CN=5) is not a good glass former (breakdown of rule 2)
MgO (rocksalt structure , CN=6) is not a good glass former (breakdown of rule 2)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 Zachariasens theory applied and revisited

Extension to modified oxides

Warren and Bisco, 1930s

Modified Random Network

Greaves, 1989

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Why do glasses form ?

Zachariasen's rules do not consider at all modified oxides or multicomponent systems, or

even non-oxide glasses

Sulfur or selenium based glasses do form quite easily glasses

Extensive studies in composition have led to the definition of the glass forming region

Gunasekera et al. JPC B (2013) Saienga and Martin, JNCS (2008)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Why do glasses form ?

Competition between crystalline growth and cooling the melt.

Good glass formers (q=dT/dt~10-3 K/s), Bad glass formers (q=dT/dt>106K/s)

Quantitative estimate through time-temperature-transformation (TTTT) curves

How much time does it take at any one temperature for a given fraction of the
liquid to transform (nucleate and grow) into a crystal?

Critical cooling rate Rc=

the nose of the TTT curve

See: http://www.lehigh.edu/imi/docs_LL/GCC/Lecture_2_Martin.pdf R. Busch, JOM (2000)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 Why do glasses form ?

Richet (2006): Glasses form more

easily in eutectics because these
freezing depressions bring the liquid
to higher viscosities at lower
temperatures

Correlation between critical cooling

rate and viscosity along the liquidus
branche.

Richet et al.
Chem. Geol. 2006

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY

Measuring glass-forming stability from DSC

A calorimetric mesurement of the glass transition Tg

and the recrystallization temperature Tx provides a
measure of the stability of a glass.

Gunti and Asokan, JNCS 2010

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Glass families: Network formers

Chalcogenide network formers: mixing of Group III, IV, V ,VII

with (S, Se, Te). Off-stoichiometric glasses.
e.g. SixSe1-x, PxS1-x, GexTe1-x, Ge-As-Se,

Cross-linking into a base Se (Tg=316 K) or S (Tg=245 K) glass

Variety of connectivity motifs and chemical bonding:
corner-sharing (CS), edge-sharing (ES) and homopolar bonding

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Glass families: Network formers

Oxide network formers at stoichiometric compositions

e.g. SiO2 (silica), GeO2 (germania), B2O3, P2O5
Corner-sharing (CS) connections of base structural motifs (triangles, tetrahedra).

SiO2, GeO2

B2O3

Th. Charpentier et al. JPC 2009 A. Wong et al. PCCP, 2009

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Glass families: Binary, ternary and multicomponent modified glasses

Network former (usually at stoichiometric composition) and modifiers

Modifier (alkali or alkaline earth oxide or chalcogenide)

e.g. SiO2 -Na2O-CaO-Al2O3-Fe2O3- Magmas

(1-x-y)SiO2 xNa2O-yCaO Window glass
(1-x)GeS2 xLi2S Solid electrolytes
60 SiO2-36 CaO-4 P2O5 Bioglass
(1-x)GeSe3-xAg Programmable Memories

Physical/chemical/mechanical/optical/etc. properties heavily depend on the

composition.

Le nombre de compositions possibles est quasiment infini

J.Cl. Lehmann, St. Gobain chair

Huge space for improvments, new compositions, new glasses, new studies

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Glass families: Binary, ternary and multicomponent modified glasses

Effect on structure
In tetrahedral systems, modifiers (alkali or NBO
alkaline earth oxide or chalcogenide)
disrupt the basic network former structure
and replace bridging oxygens (BO) by non-
bridging oxygen (NBO). BO

Analogue in sulphides (BS and NBS).

The population Qn of tetrahedra having n=4

3, 2, 1 0 BOs is changing with modifier
composition, temperature or pressure.

Q3 Q2 Q1 W.J. Malfait et al., JNCS 353, 4029 (2007).

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 C) GLASS FAMILIES AND GLASS FORMING ABILITY
Glass families: Binary, ternary and multicomponent modified glasses

Charge compensation
In other systems (borates or aluminates),
modifiers can either
depolymerize the base network (NBOs)
increase the coordination number of
some species (B3-> B4 in borates)
because of charge compensation.

Analogue in sulphides (BS and NBS).

Fraction of Al5
The species population also is changing
with modifier composition, temperature or
pressure.

SiO2-Al2O3-CaO
Neuville et al. Chem. Geol. 2006

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE

Among various methods

Neutron or X-ray diffraction

access to information on short and intermediate range order
access to dynamics (vibrational density of states)

Raman or infrared spectroscopy

access to access to information on vibrational dynamics and structure

NMR
probe of structure and connectivity

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE

Neutron or X-ray diffraction

Measure of a scattering cross-section given by:

Neutrons: FN(Q)

X-ray: IX(Q)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 An isotopic neutron diffraction expriment will give access to the partials S(Q)
different coherent neutron scattering lengths

Anomalous X-ray scattering (change in fi(Q)) will give access to the partials S(Q)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Neutron or X-ray diffraction: Isotopic substitution

Faber-Ziman formalism : element specific

e.g. formation of skydivers illustrates order on
an interlediate length scale
SRed-Red(k)
SRed-Green (k)
SGreen-Green(k)
Each diver has simple set of rules for bonding to
the next but there is sufficiently flexibility for
different patterns of ordering to be created on the
scale of a few body lengths.
Salmon, Nature (2002).
e.g. SGeGe(k), SGe-Se(k) in GeSe2

Bhatia-Thornton formalism (linear

combination of FZ)
SNumber-Number(k) topology-b
SConcentration-concentration(k) chemical ordering-c
SNumber-concentration(k)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Neutron or X-ray diffraction: Back to real space

Total distribution function

T(r)=4n0rg(r)

Access to total or partial

functions

Neutrons X-ray
C. Benmore, Nxschool 2009

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Raman and Infrared spectroscopy

Infrared (IR) : direct coupling between elctric field E and instaneous dipolar
momentum p. Interaction energy given by: r r
W = p.E

Happens each time the charges are displaced (modification of the dipolar momentum)

Raman : coupling between E and the atomic polarizability (ability of the electron
coloud to be deformed under the action of E).
r rr r
Induced electric polarisation : P = E
(rank 2 tensor possible anisotropies)

Expression of is modulated
r r by the atomic vibrations (harmonic
approximation) q = q 0 cos 2 v t
r r
Incident field : E = E0 cos 2 0t
r rr r
which creates an induced dipole : P = E
[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE
 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Raman spectroscopy

Assume a vibrational mode such as 2 v r

Polarisability is given by :
1
= + + +
2
1
= + + +
2
At the first order, one can write: = +

and:
= = + = 2 + 2 v 2

1 1
= 2 + 2 v + 2 + v
2 2
Rayleigh Stokes
Anti-Stokes
Can be extended to higher ordrer
[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE 23
 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Raman spectroscopy

CCl4

Atomic and molecular vibrations (mass and force constants between ions) lead to the Raman
spectrum
Mass = chemical composition
Force constant = structure , constraints (bonds, angles)

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Raman spectroscopy and probe of glass structure

Raman lines in glasses provide different informations on structure

Line position: Connection with

Chemical species;
symetries vibrational density of states
Shift:
constraints,
temperature
concentration
Intensite:

With C() coupling function

and n(,T) the Bose-Einstein statistics
(vibrations=bosons)
1
Wave number (cm-1) ! #$
Width : " %& 1
Structural disorder

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Raman spectroscopy and probe of glass structure

Raman spectra of vitreous SiO2 (silica)

Boson peak at low frequency.
Main band
Main band at 440 cm-1 together
440cm-1
with a shoulder peak D2
D2
490cm-1
Shoulder at D1 (600 cm-1).
Intensit Raman

Boson peak
-1
50cm

D1 800cm-1 HF bands
600cm-1
1050cm-1 et 1200cm-1 T. Deschamps, 2009

Frquence cm-1

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Raman spectroscopy and probe of glass structure

Measure of spectral changes with composition

Gaussian decomposition
of the main lines

Analysis of structural changes

Edge-sharing mode intensity

Bhosle et al. IJAGS 2012

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Nuclear Magnetic Resonance

NMR active nuclei possess a magnetic dipole moment

13C, 31P, 29Si, 77Se,

The coupling of this magnetic dipole moment to an applied external

magnetic fields as well as the magnetic dipole moments of neighboring
nuclei, is used in the solid state NMR experiment to characterize structure
and bonding arrangments.

Excitation frequency 0 of the magnetic field (adapted to match energy

differences between magnetic dipole excitations)

NM Resonance (Larmor) frequency at 0=B with gyromagnetic ratio

(unique for a given nucleus). Allows to adapt B in order to probe 0 for a
given isotope.

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Nuclear Magnetic Resonance
Resonance frequency is also depending on the magnetic interaction with neighboring
atoms (nuclei)
Intramolecular interaction
Intermolecular interaction (dipolar interaction)
+electrical (quadrupolar moment, Spin angular momentum>1/2)

Difference in frequency induced by different magnetic/electric atomic environment

Involves an effective screening constant (dia- or para-magnetism, etc.)

' (1 ))

The chemical shift (in ppm) is defined with

respect to an arbitrary reference (crystal, ref.
compound).
' , -
+ . 100
Evolution of the chemical shift with concentration,
temperature, pressure provides information about
structure.
[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE
 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Nuclear Magnetic Resonance and structure

1. In silicates, identification of specific NMR signals depending on the local

structure (e.g. Qn speciation), and deconvolution (estimation of population)

2. Deconvolution of a given spectra and estimation of the Qn population

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE
Nuclear Magnetic Resonance and structure D. Kasemann et al. JPC B (2013)

3. Characterization of structure and evolution

Ge-Se
with chemical composition

Qn speciation in lead silicates

Massiot et al., JNCS 2008

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

 D) EXPERIMENTAL CHARACTERIZATION OF STRUCTURE

Having more interest in experimental characterization ?

Go to: http://www.lehigh.edu/imi/GlassStructureCourse.htm

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE

Conclusion: Before getting started, which theory for which kind of experiments ?
Experiments Theory
Technique Quantity MD DFT Reverse Monte Carlo Bond models
SRO, MRO
Neutron S(k), g(r),
X-ray Partial gij(r)
VDOS
EXAFS, SRO
XANES Spectra
Anomalous SRO, MRO, partial
X-ray S(k), g(r)
Raman
Vibrational
Infrared
NMR Speciation, SRO,
MRO
XPS eDOS, SRO
Dynamics D/ viscosity
= Easy and/or excellent agreement = Medium difficulty and/or fair agreement
= Difficult and/or poor agreement
 Home reading
80 years of random networks, M.F. Thorpe and A. Wright, Physica Status Solidi (2013).
pdf file

The atomic arrangement in glass, W.H. Zachariasen, J. Am. Chem. Soc (1932).
pdf file

[email protected] Atomic modeling of glass LECTURE 2 STRUCTURE