New Journal of Glass and Ceramics, 2011, 1, 28-33
doi:10.4236/njgc.2011.12005 Published Online July 2011 (
Copyright © 2011 SciRes. NGJC
Evaluation of Non Crystalline Phase in AZS
Refractories b y XRD Methods
M. S. Conconi1,2,3, N. M. Rendtorff1,3,4, E. F. Agli e tti1,4,5
1CETMIC (Centro de Tecnología de Recursos Minerales y Cerámica, (CIC-CONICET-CCT La Plata)), M. B. Gonnet, Argentina;
2Facultad de Ingeniería de la Universidad Nacional de La Plata. Argentina; 3CIC-PBA, Buenos Aires Argentina; 4Facultad de
Ciencias Exactas de la Universidad N acional de La Plata. Argentin a; 5CONICET La Plata, Argent ina.
Received Ma y 20th, 2011; Revised J une 17th, 2011; Accepted June 24th, 2011.
The relation between the atomic structure and the macroscopic properties and behaviors of a material constitute one of
the obje c tiv e s of the materials sc ie nce, particularly in the design and development of ceramic materials. Crystalline and
non crystalline phases together with pores, grain boundaries, etc. affect mechanical and fracture prop-erties as we ll a s
chemical resistance and electric properties. These aspects will be bonded to the raw materials chosen and the whole
processing route. In glass industry, although there are other electrofused refractories such as the alumina ones used in
the feeding of the fusion kilns, probably the most used refractories in contact with the melted glass are electrofused ma-
terials that belong to the Al2O3-SiO2-ZrO2 system commonly named AZS. Exceptionally for refractory materials the
amount of the glassy phase in a AZS material is important and appreciable; and makes them particularly adequate for
containing fussed glass. The glass proportion will define much of their prop-erties and behaviors. In the present work
the results of the non crystalline phase quantification of two samples of commercial AZS materials are presented and
compared. These were obtained by three different methods using in the X ray powder diffraction (XRD) techniques. The
first method consists in the linear interpolation of the base lines of the diffractograms compared to the amorphous silica
and th e fully crystalline quartz. The other two methods are based in the application of the Rietveld method. One is the
internal standard method with quartz as fully crystalline standard and the other one consist in the inclusion of the glas-
sy phase to the refinement with a structural model that can be understood as the widening of the peaks consequence of
an extreme decrease in the crystallite size of a quartz phase. The three methods showed equivalent results (with d iffer-
ences less than 3%) for the two samples and demonstrated that are adequate for the quantification of the non crystalline
phase in this kind of materials.
Keywords: XRD, Rietveld Method, Non Crystalline Pha se, Refractories
1. Introduction
In glass industry, although there are other electrofused
refractories such as the alumina ones used in the feeding
of the fusion kilns, probably the most used refractories in
contact with the melted glass are electrofused materials
that belong to the Al2O3-SiO2-Zr O 2 system commonly
named AZ S .
In gla ss ind ustr y, altho ugh t here are o ther electrofused
refractories such as the alumina ones used in [1,2]. T hese
materials have influenced drastically the quality levels
and productivity of the glass fabrication processes [3].
The first advance achieves by these materials was the
improvement in the corrosion resistance, increasing the
life of the kilns, together with the quality of the
processed glass. Other important progresses accomplish
were the improvement in the mou nting time and d ecrease
in the devitrification cords and lower amount of bubbles
produced by the “blistering” process [4-6].
Exceptionally for refractory materials the amount of
the glassy phase in a ASZ material is important and ap-
preciable; The weight proportions is around 20% and
makes them particularly adequate for containing fussed
glass. The glass proportion will define much of their
properties and behaviors.
To validate and compare three quantification methods
for the non crystalline phase of an electrofused refractory
material is the principal objective of the present wo rk. In
the materials field the use of the X ray diffraction tech-
nique fo r the non c r yst al li ne or a morp hous fra ct io n q ua n-
Eva lua t ion of Non Crystalline Phase in AZ S Refr actories by XRD Methods
Copyright © 2011 SciRes. NGJC
tification is a permanent challenge that has been studied
from diverse forms. Three of them are compared in the
present wor k.
The typical composition of an AZS material of the
normal filling cast type, from the supplier information is
sho wn in Table 1 .
Ohlberg [7] developed a method for determining the
crystallinity percentage (C%) in partially devitrified
glasses, by the interpolation of the base line of the dif-
fractogram between the corresponding to amorphous
silica and fully crystalline quartz.
( )
( )
% 100g
= ×
Where Ig, I m y Ic are the diffractogram intensities at 2θ
= 22.5 corresponding to a sample 100% glass
(amorphous silica), the partiall y crystalline phase (prob-
lem sample) and the 100% crystalline standard (quartz)
This equation commonly utilized for determining
crystallinity i n vitro-ceramic materials in a wide range of
proportions [8,9], it could be valid to assume that the
amorphous or non crystalline (NCO%) proportion can be
obtained from the following e quation defining NCOh% as
a compleme nt of the crystallinity (NC% + C% = 100).
( )
( )
( )
( )
% 100% 100100
gc gc
=−=× =×
The Rietveld method [10,11] has demonstrated to be
an effective tool for quantitative phase analysis in di-
verse materials [12,13]. The quantitative analysis is car-
ried out from the scale factors refined for each phase (Si)
according to the following equation:
( )/
( )/
i ii
ip pp
Table 1. Typical compositions of an electrofused AZS re-
fract ory .
compositiona (% wt.) Mi neralo g ical
composition (% wt.)
ZrO2 + HfO2 35.4 m-ZrO2 33.0
Al2O3 48.0 Al 2O3 47.0
SiO2 15.2 Amorphous phase 20.0
Na2O 1.5
TiO2 0.04
Fe2O3 0.04
a. -azs.a spx
Where Wi is the weight fraction of the i-phase over all
the present phases. Si, Zi, Mi, Vi and τi are the scale
factor, the number of molecules per unit cell, the mo
lecular weight, the unit cell volume and the mass-absor-
ption correction factor of the particles for the i-phase,
respect tively.
In the Rietveld analysis, the crystalline structure of
each phase in the sample should be known. Hence this
method does not allow including the amorphous or
non-crystalline phases. However several authors had put
into practice the quantification of these phases using the
Rietveld refi nement i n efficie nt way. De la T orre applied
the method for samples with the aggregate of a fully
crystalline (100%) internal standard in a known propor-
tion, and determined the experimental conditions which
affect the uncertainty of the amorphous phase determina-
tion using different internal s tandards [14].
Le Bail demonstrated that it is possible to include the
silica glass in the Rietveld refinement through a struc-
tural model with crystalline defects [15]. Lutterotti [16]
applied Le Bail method for the silica glass introducing
defects from the crystal size for reproducing the peak
widening; verifying this method for standard samples of
quartz and amorphous silica, after he applied it to sani-
tary ceramic and to a AZS refractory. Finally Ward [17]
compared two Rietveld methods in flaying ashes. The
first one with the internal standard aggregate in known
proportion and other one introducing the amorphous
phase in the refinement program through the incorpora-
tion of experimental standards of non crystalline phases
like meta-caolin or tr idimite.
Mechanical mixtures of crystalline and non crystalline
had been commonly used as standards for studying the
efficiency of non c ryst all ine quantifi cation methods.
In electrofused or sintered materials, these mixtures
are not the most adequate, due to the fact that the phase
distribution in the standards differs from the actual stu-
died materials.
In the first case crystalline and non crystalline par-
ticles are clearly differentiated and produce different
diffractions compared to the produced in samples with
particles where both type of phases are together in the
same particle. In consequence the comparison of diverse
metho ds with sa mples with u nknown a morp hous cont ent
will allow validating them.
In the present work the results of the non crystalline
phase quantification of two samples of commercial AZS
materials are presented and compared. These were ob-
tained by three different methods based in the X ray
powder diffraction (XRD).
The redefinition of the Ohlberg equation (Equation 2)
was used for the first method. Milled quartz (SiO2) was
Eva lua t ion of Non Crystalline Phase in AZS Refracto r ies by XRD M ethods
Copyright © 2011 SciRes. NGJC
used as fully crystalline standard for the first Rietveld
refinement method. Finally the Le Bail model based me-
thod was applied with the amorphous phase incorporated
as a nanocrystalline material with a ß-Carnegieite struc-
2. Experimental Procedures
The analyzed material consisted in a monolithic electro-
fuse d commercial AZS refractory normal filling type
(AZS ER 1681 RN, Saint-Gobain SEFPRO, Italy). Par-
ticularly two samples (AZS1 y AZS2) of the material
were studied coming from different blocks. For the anal-
ysis samples were milled in Agatha mortar up to mesh
The chemical analysis of the samples was carried out
by Atomic Emission Spectroscopy by inductive coupled
plasma (Varian Vista AX CCD Simultaneous ICP-AES)
with the exception of the zirconium which was done by
X ray Fl uorescence (Shimadzu EDX800HS).
For the amorpho us phase characterization Silicon dio-
xide (SiO2) powder was used as standard (Carlo Erba
RPE) for obtaining Ig in the Ohlberg method and for
refining the pure glassy phase in the Rietveld Method.
Also 15%wt. of milled quartz was used as internal stan-
dard aggregate before the Rietveld quantification. This
was chosen because it presents a similar absorption coef-
ficient to the sample [14]. For obtaining the Ic T he same
crystalline quartz was used.
Materials were analyzed by XRD (Philips 3020
equipment with Cu Kα radiation in Ni filter at 40 kV to
20 mA). D ifractograms were carried out between 10 and
70 in 2θ with 0.04 steps of 3 seconds.
The powder XRD patterns were analyzed with the
program FullProf [18], which is a multipurpose pro-
file-fitting program, including Rietveld refinement. The
starting crystallographic data for each phase were ex-
tracted from the literature.if any non crystalline phase is
present in the sample when the Rietveld refinement done,
the internal standard content would be overestimated.
The percentage of amorphous phase in the sample with-
out the aggregated standard can be calculated using the
following equation [14]:
(1/ )
%10 %
IS s
= ×
Where NCIS% is the non cr ystalline content by the inter-
nal standard method, WS is the internal standard propor-
tion aggregated (%) and RS is the internal standard eva-
luated by the Rietveld method.
For obtaining the actual phase content of each present
phase they should be corrected by the amorphous phase
The only refined parameter in the Le Bail model re-
finement was the scale parameter. For the other phases
scale factor was accompanied by the cell parameters, and
the rest of the parameters which describe the profile. The
background was calculated from the interpolation of
several 2θ: intensity pairs. Moreover the background was
not refined between 5˚ and 45˚, while in the rest of the
diffractogram they were refined with the other of the
Befo re intro ducing the no n crystal line p hase i n the re-
finement of the studied samples, the pure amorphous
silica was analyzed for determining the crystalline and
profile parameters.
3. Results and Discussion
The results of the chemical analysis of the samples are
presented in Table 2. The ZrO2 content was calculated
from the ele mental Zr content obtained by XRF.
In the XRD test Al2O3 together with monoclinic and te-
tragonal zirconia were detected. There were not detected
any Silicon (Si) containing crystalline phase, evidencing
an important silica rich ( 70%) non crystalline phase.
This fact supports the assumption of approximating the
non crystalline phase of these materials wi th silica glass.
3.1. Ohlberg method
In order to apply Ohlberg equation (Equation 2) the cor-
responding intensities of the diffractograms at 2θ = 22.5
for both samples AZS1 and AZS2. Both samples pre-
sented almost ide ntical intensi ties, in Figure 1 a detail of
the superposed diffractograms for AZS1 sample, the
amorphous silica and crystalline quartz between 15˚ and
35˚ is shown and i n Table 3 the results of the amorphous
quantification of both samples are revealed.
3.2. Internal Standard Method
In Figure 2 the diffractogram with its corresponding
refinement curve is shown for sample AZS2 with the
15%wt. quartz aggregate. There it can be observed the
experimental profile (dots) and the theoretical profile
(continuous), the corresponding positions of the diffrac-
tion lines of each phase (alumina, monoclinic zirconia,
quartz and tetragonal zirconia respectively) are expose in
vertical bars, finally the difference between the observed
profile and the theoretical pr ofile is shown in the base of
the graph.
After the Rietveld refinement the quartz evaluated
content was 19.1%wt. in AZS1 and 19.2%wt in AZS2.
The actual contents of non crystalline and crystalline
phases by this method are shown in Table 4. T he evalu-
ated crystalline and non crystalline content for the two
different samples are equivalent.
Table 2. Chemical composition of the studied materials.
Eva lua t ion of Non Crystalline Phase in AZ S Refr actories by XRD Methods
Copyright © 2011 SciRes. NGJC
Samp le
SiO2 15.4 14.7
Al2O3 45.9 45.9
Fe2O3 0.33 0.42
CaO 0.10 0.12
MgO 0.02 0.02
Na2O 1.29 1.28
K2O 0.03 0.03
TiO2 0.08 0.04
ZrO2 + HfO2 32.4 3 2.3
a:ICP and XRF results
Table 3. Non crystalline content evaluated by the Ohlberg
me thod (Equa ti o n 2).
Sample NCOh%
AZS1 23.3
AZS2 23.2
Table 4. Quantitative analysis results from the Rietveld
Me thod, with inte rnal Standard.
Al2O3 46.5 47.3
m-ZrO2 27.8 26.8
t-ZrO2 1 1
NCIS% 24.7 24.9
Table 5. Quantitative analysis results from the Rietveld
Me thod by Le Bail model.
Al2O3 48.1 42.9
m-ZrO2 28.0 30.3
t-ZrO2 1 1
NCLB% 22.9 25.8
Figure 3 presents de Rietveld refinement figure using
the Le Bail model in sample AZS1 made with the me-
thod described before. Diffraction lines correspond to:
Alumina, Monoclinic Zirconia, amorphous silica and
tetragonal zirconia respectively. Non crystalline and
crystalline phase contents are presented in Table 5. The
m-ZrO 2 and non crystalline content for the two different
samples are almost equivalent2. The alumina content in
AZS1 is higher than the one evaluated in the other sam-
Figure 1. Crystalline quartz, amorphous silica and AZS1
sample diff ract ogra m between 1 5˚ - 35 ˚.
Figure 2. Rietveld refine ment of sample AZS2 w ith crystal-
line quartz as int ernal standard.
Figure 3. Rietveld refinement of sample AZS1 with the Le
Bail model.
Eva lua t ion of Non Crystalline Phase in AZS Refracto r ies by XRD M ethods
Copyright © 2011 SciRes. NGJC
Figure 4. Co mpari so n of t he non cry st alli ne c ont ent eval ua-
tion by the three appl ied methods.
4. Summary
The quantification results are compared in Figure 4 (bar
chart). Although the three methods are based in com-
pletely different principle, their re sults are e quivalent,
with differences below 3% for the studied material from
two di fferent samples , showi ng that t he three models are
adequate for the studied system, moreover the results
match with the results provide by the material supplier
(Table 1). In fact the results for the three applied me-
thods are slightl y hi g her .
Although the simplicity of Ohlberg method, it can be
applied only for materials that do not present diffraction
lines in 2θ = 22.5, and this method do not provides in-
formation ab out the other crystalline phases.
A complete phase quantification (crystalline and non
crystalline) can be carried out by both Rietveld refine-
ment based methods, but the Lebail model is recom-
mendable because it is not necessary to contaminate the
sample with the addition of the internal standard and it
could be easily incorporated to a routinely Rietveld
phase quantitative analysis without any increase in the
number of X ray diffractograms.
[1] G. Duvierre, E. Sertain and A. Rebert, “Advantages of
Using High Zirconia Refractories in Lead Crystal Glass
Electricfurnaces,” Glass Technology, Vol. 34, No. 5,
1993, pp . 181-186.
[2] P. C. Ratto, “Réfractaires Ele ctrofondus du Systeme
AZS: Différentes Méthodes de Fabrication Oxydantes et
Leurs Impacts sur le Comportement du Réfractaire en
Service,Verre, Vol. 8, No. 3, 2002, pp. 22-27.
[3] E. Lataste, “Comportement Mecanique et Endommage-
ment de Refractaires Electrofondus sous Sollicitation
Thermomecanique,” Ph.D. Dissertation, INSA de Lyon,
[4] S. Yamamura, M. Kitano and Y. Kakimoto, “An Inte-
grated Approach to Optimum Furnace Design,Glass In-
ternational, Vol. 30, No. 1, 2007, pp. 40-41.
[5] J. Zborowski, “Some Aspects of Characterization of the
Refractories for Glass Contact,Proceedings of the Uni-
fied International Technical Conference on Refractories:
the 9th Biennial Worldwide Congress on Refractories,
2006, pp . 690-694.
[6] S. M. Winder, K. R. Selkregg an d A. Gupta, “Update on
Selection of Refractories for Oxy-Fuel Glass-Melting
Service,Ceramic Engineering and Science Proceedings,
Vol. 2 0, No . 1, 1999, pp. 81-105.
[7] S. M. Ohlberg and D. W. Strickler, “Determination of
Percent Crystallinity of Partial Devitrified Glass by
X-Ray Diffraction,Journal of the American Ceramic
Soci ety, Vol. 45, N o. 4, 1962, pp.170-171.
[8] J. P. Willams, G. B. Carrier, H. J. Holland and F. J.
Farnco mb, “The Determination of the Crystalline Content
of Gl a s s -Cer amics,Journal of Materials Science, Vol. 2,
No. 6, 1967, pp. 513-520. doi:10.1007/BF00752217
[9] S. Morimoto, “Phase Separation and Crystallization in the
System SiO2-Al2O3-P2O5-B2O3-Na2O Glasses,Journal
of Non-Crystalline Solids, Vol. 352, No. 8, 2006, pp.
756-760. doi:10.1016/j.jnoncrysol.2006.02.007
[10] H. M. Rietveld, “A Profile Refinement Method for Nuc-
lear and Magnetic Structures ,Journal of Applied Crys-
tallography, Vol. 2, No. 2, 19 69 , pp. 65 -71.
[11] R. A. Young, “The Rietveld Method,” International Un-
ion Crystallography, Oxford University Press, Oxford,
[12] D. L. Bish and S. Howard, “Quantitative Phase Analysis
Using the Rietveld Method,Journal of Applied Crystal-
lography, Vol. 21, No. 2, 1988, pp. 86-91.
[13] N. V. Y. S carlet t, I. C . Mads en, L. M. D. Cranswick, T. L.
Edward Groleau, G. Stephenson, M. Aylmore and N.
Agron -Ol shina, “Outcomes of the International Union of
Crystallography Commission on Powder Diffraction
Round Robin on Quantitative Phase Analysis: Samples 2,
3, 4, Synthetic Bauxite, Natural Granodiorite and Phar-
maceuticals,Journal of Applied Crystallography, Vol.
35, No. 4, 20 02, pp. 38 3-400.
[14] A. G. De La Torre, S. Bruque and M. A. G. Aranda,
“Rietveld Quantitative Amorphous Content Analysis,
Journal of Applied Crystallography, Vol. 34, 2001, pp.
196-202. doi:10.1107/S0021889801002485
[15] A. Le Bail, “Modelling the Silica Glass Structure by the
Rietveld Method ,Journal of Non-Crystalline Solids,
Vol. 183, No. 1-2, 1995, pp. 39-42.
doi:10.1016/0022-3093(94 ) 00 66 4-4
[16] L. Lutterotti, R. Ceccato, R. Dal Maschio and E. Pagani,
“Quantitative Analysis of Silicate Glass in Ceramic Ma-
terials b y de Riet veld Met hod,Material Science Forum,
Vol. 278-28 1, 19 98, pp. 87-92.
[17] C. R. Ward and D. French, “Determination of Glass
Eva lua t ion of Non Crystalline Phase in AZ S Refr actories by XRD Methods
Copyright © 2011 SciRes. NGJC
Content and Estimation of Glass Co mposition in F ly Ash
Using Quantitative X-Ray Diffractometry,” Fuel, Vol.
85, 2006 , pp. 22 68–2277. doi:10.1016/j.fuel.2005.12.026
[18] J. Rodríguez-Carvajal, “Recent Developments of the
Program Fullprof,” Newsletter in Commission on Powder
Diffraction (IUCr), Vol. 26, 2001.