Available online at www.sciencedirect.

com

Journal of the European Ceramic Society 33 (2013) 653660

Characterization of LiZnFe crystalline phases in low temperature

ceramic glaze
Anucha Wannagon , Sitthisak Prasanphan, Samunya Sanguanpak
National Metal and Materials Technology Center, Thailand Science Park, 114 Paholyothin Rd., Klong 1, Klong Luang, Pathumthani 12120, Thailand
Received 26 August 2012; received in revised form 9 October 2012; accepted 11 October 2012
Available online 23 November 2012

Abstract
Ceramic glaze containing Li2 O and ZnO was prepared at a low firing temperature of 1100 C. Addition of 030 wt.% iron oxide content developed
brown color with a metallic sparkling effect from crystallization after soaking at 9801080 C. Using XRD, SEM/EDS and Raman microscopy the
crystalline phases were determined as lithium zinc ferrite (Lix Zn12x Fe2+x O4 where x = 0.050.20), hematite (-Fe2 O3 ) and anorthite (CaAl2 Si2 O8 ).
The most preferable metallic sparkling effect was caused by the lithium zinc ferrite phase obtained from the glaze containing 10 wt.% of iron oxide.
Thermal analysis by STA after heat treatment indicated that crystallization temperature of lithium zinc ferrite and the effective soaking temperature
depended on the iron oxide content in the glaze. The influence of excessive iron oxide content on the crystallization behavior of lithium zinc ferrite,
anorthite and hematite phases is discussed.
2012 Elsevier Ltd. All rights reserved.
Keywords: Crystalline glaze; Low firing temperature; Lithium zinc ferrite; Anorthite; Hematite

1. Introduction
Crystalline glaze is a devitrified glaze which consists of crystalline phases produced during controlled nucleation and growth
processes.1 This type of glaze is normally produced by gloss firing at high temperature above 1250 C and subsequently soaking
in the temperature range of 11001250 C. Addition of colored
metal oxides such as iron oxide can generate crystalline effects
in glazes. However the glaze usually needs to be gloss fired
in the high temperature range of 13001350 C.2 Most of the
effects from iron oxide addition result from the presence of the
crystalline groups of hematite (-Fe2 O3 ) under oxidizing conditions or magnetite (Fe2 O4 ) under reducing conditions. Anorthite
crystals (CaOAl2 O3 2SiO2 ) were reported to be formed when
additions of 1220% of Al2 O3 were made.3 Another effect of
iron oxide crystals appears at lower temperature resulting in
the so-called aventurine crystalline glaze, which is traditionally
obtained when using lead oxide at lower firing temperatures. Through the use of lead-free Na2 OB2 O3 SiO2 frit, the
iron-aventurine glaze with a crystalline effect from hematite was
reported to be obtained in the temperature range 10001300 C.4

Corresponding author. Tel.: +66 2564 6500; fax: +66 2564 6368.
E-mail address: anuchaw@mtec.or.th (A. Wannagon).

0955-2219/$ see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jeurceramsoc.2012.10.027

A low firing temperature glaze without lead oxide can be

produced by adding alkali and alkaline earth oxides such as
Na2 O, K2 O, Li2 O, CaO and MgO. The addition of Na2 O and
K2 O decreases the glaze viscosity and facilitates crystal formation. On the other hand, MgO, CaO and BaO additions increase
glaze viscosity and prevent crystal formation, hence these oxides
should be kept below a certain level otherwise they can cause
the glaze to develop a matt appearance.5 Low firing temperatures could be used when small amounts of ZnO and B2 O3 were
added. In this case, under suitable conditions, a high content
of zinc oxide can easily act as a nucleating agent and zincbased crystalline phases can be obtained. The zinc-containing
glaze is usually gloss fired at 11501300 C.6 It gives the well
known macrocrystalline glaze due to the formation of willemite
(2ZnOSiO2 ) crystals which can grow up to centimeters in size.
The effects of different gloss firing temperatures in the range
12001250 C on transformation of willemite to spherulite has
been reported.7 The extent of willemite crystal formation was
reduced when CaCO3 content increased up to 3% at 1200 C
gloss firing temperature.8 It was considered that coloring additives such as, Fe2 O3 did not have as much effect as alkali and
alkaline earth oxides on the crystallization of willemite since
crystals continued to grow in the colored matrix.3 The effect of
iron oxide content on crystallization of the glass ceramic glaze
has also been studied.9 The crystalline phases of franklinite

654

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

(ZnFe2 O4 ) and hematite (Fe2 O3 ) were obtained and their relative amounts depended on the iron content in the frit mixture.9
Glaze containing Li2 O may show different crystallization
behavior since Li2 O is a strong flux and can influence changes
in the oxidation state of Fe2 O3 in the glaze. It has been
reported that lithium metasilicate easily crystallized at low temperature in a Li2 OFe2 O3 SiO2 glass.10 LiFeSi2 O6 , LiFe5 O8
and LiFeO2 crystalline phases were formed during crystallization in CaOMgO (Li2 O, Fe2 O3 )SiO2 glasses.11 In the
LiO2 ZnOSiO2 system, it was thought that Li2 ZnSiO4 phase
developed due to the reaction of ZnO with Li2 Si2 O5 .12 A small
addition of Li2 O changes the crystallization path by precipitation of -quartz solid solution and willemite at low temperature
(800900 C).13 Doping by Li2 O also stimulated the formation
of franklinite in Fe2 O3 ZnO solidsolid interaction.14 However,
it is still not clear which phases will form when Fe2 O3 is added
to glazes containing Li2 O and ZnO.
Lead-free low temperature glaze formulations containing
Li2 O and ZnO whose properties are compatible with commercial stoneware body have been developed and applied in ceramic
manufacturing. Base glazes with gloss firing temperatures in the
range 10001100 C need to be developed in terms of color and
effect to enhance the value of glazed products. Since crystallization of Fe2 O3 possibly occurs more easily when Li2 O and ZnO
are contained in the glaze, it is of interest to study the crystalline
phases that develop at temperatures below 1100 C.
2. Materials and methods
The test glazes were produced at low firing temperature by
a crystallization technique. The base glaze was prepared from
commercial grade raw materials (the suppliers) as follows:
nepheline syenite (Cernic International Co., Ltd.), pottery
stone (C.K. Ceramics Co., Ltd.), colemanite and silica sand
(Amarin Ceramic Corp. Co., Ltd.). Chemical compositions of
the raw materials were analyzed by X-ray fluorescence (XRF,
Philips PW-2404). The Sager formula was used to calculate the
trial glaze formulation. 5 wt.% of commercial grade Li2 CO3
(Ceramic R Us Co., Ltd.) and ZnO (Cernic International Co.,
Ltd.) were used to reduce gloss firing temperature. Commercial
grade iron oxide (Amarin Ceramic Corp. Co., Ltd.) 530 wt.%
was added to develop crystalline effects in the glazes which are
referred as Fe5Fe30 formulas. The raw material mixture was
mixed with 80% water and milled by rapid milling for 10 min.
The glaze slip was sieved and 0.3 wt.% of commercial grade
dispersive agent Na2 SiO3 (Cernic International Co., Ltd.) was
added. The density was adjusted to 1.5 g/cm3 by water. The
glaze slip was applied to the biscuit test piece (stoneware body,
4.55 cm) by dipping. The test piece was fired at a heating
rate of 2 C/min to 1100 C followed by soaking for 1 h. It was
cooled down at a cooling rate of 0.5 C/min to 1080 C, soaked
for 3 h, and consequently cooled down at a rate of 1 C/min
to 980 C, then soaked for 1 h. The crystalline glaze materials
were obtained after cooling down to room temperature. The
experimental procedure is shown in Fig. 1.
The crystalline phases formed in each glaze were identified
by X-ray diffraction (XRD, JEOL JDX-3530, Cu-K radiations,

Fig. 1. Preparation of the crystalline glaze.

30 kV, 40 mA, scan rate 0.04 s1 ). Microstructures and elemental compositions were examined by scanning electron
microscopy (SEM, Hitachi S-3400N, at 20 kV, WD = 910 mm)
and energy dispersive X-ray spectrometry (EDS, EDAX Genesis
6.04, live time 100, dead time 1520%). Crystallization behavior
was observed by simultaneous thermal analysis (STA, Netzsch
STA 449C, heating rate 10 K/min). XRD and Raman microscopy
(Senterra Dispersive Raman Microscope, Bruker Optics,
785 nm, 25 mW) were used to study the crystallization changes
after the heat treatment at temperature range 8201100 C.
3. Results and discussion
After the gloss firing at 1100 C, the glazes have yellowish to brown and dark brown color with increase in iron oxide
content. Crystalline effects were found in all glazes in which
the iron oxide content was 530 wt.% (Fe5Fe30) after soaking
at 1080 and 980 C as shown in Fig. 2. There was a small crystalline effect in the glaze when the iron oxide content was 5 wt.%
(Fe5). The degree of crystallinity increased with an outstanding sprinkling metallic effect when iron oxide contents were 10
and 15 wt.% (Fe10Fe15). However this effect was decreased
as the content of iron oxide increased to 20 wt.% and above
(Fe20Fe30).
The XRD patterns in Fig. 3 showed that the main crystalline phase was lithium zinc ferrite (Lix Zn12x Fe2+x O4 where
x = 0.050.20) when the iron oxide content was 515 wt.%
(Fe5Fe15) but the peak intensity of this phase was decreased
when the iron oxide content was increased. The main phases
became anorthite (CaAl2 Si2 O8 ) and hematite (-Fe2 O3 ) when
the iron oxide content was more than 20 wt.% (Fe20Fe30).
Back scattered images from SEM in Fig. 4 showed equiaxed
shaped crystals in bright contrast and rod shaped crystals, the
equiaxed crystals being observed in every sample. The rod
shaped crystals and some smaller particles were observed when
the amount of the equiaxed crystals decreased especially in

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

655

Fig. 2. Appearance of the Fe5Fe30 crystalline glazes.

Fig. 3. XRD patterns of the Fe5Fe30 crystalline glazes.

the Fe20Fe30 glazes. The EDS analysis results presented in

Table 1 showed that the equiaxed crystals were rich in Fe and Zn,
whereas the rod-like crystals were rich in Si, Al and Ca, with the
smaller particles being rich in Fe. Although Li was not detectable
by EDS, using information from the XRD study, the equiaxed
shaped, rod shaped crystals and the particles were assumed to
be the crystalline phases of lithium zinc ferrite, anorthite and
hematite, respectively. The results indicated that the Fe10 glaze
had the most attractive sparkling metallic effect, which was due
to the crystalline effect of lithium zinc ferrite. The rod shaped
anorthite crystals were predominant in the Fe20Fe30 glazes and
were surrounded by particles of hematite the amount of which
was increased with iron oxide content.
Raman microscopy was used to study the lithium zinc ferrite, anorthite and hematite phases in the Fe10Fe30 crystalline
glazes. Raman spectrum with an optical image of the Fe10 crystalline glaze is shown in Fig. 5. The multi-shaped group of
crystals on the Fe10 surface shows a sharp peak at 1366 cm1
as well as the peaks at 663, 467, 358 cm1 . The strong peak at
high frequency corresponds to hematite.14 The 3 peaks pattern
at lower frequency is similar to the spectrum of spinel ferrite
ZnFe2 O4 as reported by Zhou et al.15 and Wang et al.16
Raman spectra of the Fe30 crystalline glaze are shown in
Fig. 6(a) and (b). In Fig. 6(a), the spectrum from small particles showed sharp peaks at 225, 292, 409, 498, 610 cm1 and
a broad peak at 1323 cm1 which was defined as the spectrum
of hematite. It matched well with the spectrum for hematite
reported by Faria et al.17 Anorthite crystallized from glass

656

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

Table 1
Element composition of crystals in the Fe5Fe30 glazes after fired at 1100 C (wt.%).
Glaze

Crystal

Na

Mg

Al

Si

Ca

Fe

Zn

Fe5
Fe10
Fe15
Fe20

Equiaxed
Equiaxed
Equiaxed
Equiaxed
Rod
Particle
Equiaxed
Rod
Particle
Equiaxed
Rod
Particle

4.47
3.69
4.88
3.87
36.52
26.24
4.76
31.60
17.53
4.93
32.44
19.06

0.16
0.20
0.19
0.15
4.74
0.68
0.21
4.86
1.59
0.24
4.91
0.57

2.50
3.58
2.47
2.55
0.25
0.39
1.24
0.27
1.15
1.93
0.17
1.13

11.83
11.63
8.00
5.69
13.91
4.53
10.20
13.34
3.31
6.78
14.10
4.93

8.16
4.95
1.45
0.83
33.78
8.55
21.92
36.55
5.61
8.51
36.42
8.05

0.82
0.61
0.15
0.11
1.18
1.05
1.40
2.44
0.76
0.84
2.73
0.41

0.55
0.51
0.26
0.20
5.26
1.14
1.94
4.82
1.03
1.91
4.48
0.93

41.10
45.47
52.73
58.49
3.42
52.88
38.19
4.66
67.94
48.82
3.81
50.05

30.41
29.36
29.87
28.11
0.93
4.54
20.14
1.46
1.07
26.04
0.95
14.86

Fe25

Fe30

Fig. 4. SEM back scattered electron images of the Fe5Fe30 crystalline glazes.

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

657

Fig. 5. Raman spectrum and image at analyzed area of the Fe10 crystalline glaze.

should provide remarkable Raman peaks at around 450520,

8001200 cm1 .18 However, Raman spectrum at the rod shaped
crystals from the Fe30 crystalline glaze present in Fig. 6(b),
showed peaks at 349, 480, 651, 1024 and 1359 cm1 . Due to
limitations in spot size the spectrum could not be clearly identified since the rod shaped crystals were surrounded by smaller

particles. It was assumed that the observed spectrum was the

mixed spectra of anorthite and hematite.
The crystallization temperature and development of lithium
zinc ferrite in the Fe10 glaze after heat treatments were of
interest. Thermal analysis results by STA for the Fe10 glaze
(Fig. 7) showed some exothermic changes in the temperature

Fig. 6. Raman spectra and images at analyzed areas of the Fe30 crystalline glaze.

658

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

Fig. 7. DTA curve of the Fe10 glaze.

Fig. 10. DTA curve of the Fe30 glaze.

Fig. 8. XRD pattern of the Fe10 glaze after heat treatments.

range 820920 C. Then two samples of the Fe10 glaze were

freshly prepared. One was fired at 820 C and the other at
920 C, for each sample the heating rate up to the firing temperature was 2 C/min. The crystallization changes after the
heat treatments were followed by XRD and SEM, as shown
in Figs. 8 and 9, respectively. In Fig. 8, the XRD pattern of the
Fe10 glaze after heat treatment at 820 C showed the presence

of anorthoclase ((Na,K)AlSi3 O8 ), franklinite (ZnFe2 O4 ) and

hematite (-Fe2 O3 ). Franklinite and hematite were not detected
but lithium zinc ferrite (Lix Zn12x Fe2+x O4 where x = 0.050.20)
was found after the heat treatment at 920 C.
SEM micrographs of Fe10 glaze after heat treatment at 820
and 920 C are shown in Fig. 9 revealing the presence of multishaped groups of crystals and rod shaped crystals. Taking into
account the results of EDS and XRD, it can be assumed that the
multi-shaped groups of crystals formed at 820 C were franklinite and that this changed to lithium zinc ferrite after heat
treatment at 920 C, while the rod shaped crystals remained
as anorthoclase at both temperatures. The DTA curve of Fe10
glaze (Fig. 7) showed an exothermic peak at 860 C, which could
correspond to the transformation of franklinite to lithium zinc
ferrite. It has been suggested by El-Shobaky et al.14 that franklinite in Zn/Fe compound can be formed at about 600 C. Heating
it to 600800 C leads the formation of non-stoichiometric zinc
oxide that contains Zn+ ions, which may be partially substituted
by Li+ resulting in lithium zinc ferrite (Lix Zny Fez O4 ) as found
in the present work.
The DTA curve for the Fe30 glaze is shown in Fig. 10.
The development of lithium zinc ferrite, anorthite and hematite

Fig. 9. SEM images of the Fe10 glaze after heat treatments.

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

Fig. 11. XRD pattern of the Fe30 glaze after heat treatments.

phases was studied after heat treatment in the temperature

range 9001100 C. After heat treatment, the sample turned to
brown or dark brown color. In Fig. 11, the XRD patterns of
the Fe30 glaze after heat treatment at 900 C showed anorthite
(CaAl2 Si2 O8 ), franklinite (ZnFe2 O4 ) and hematite. Lithium
zinc ferrite (Lix Zn12x Fe2+x O4 where x = 0.050.20) was found
after heat treatment at 9501000 C but hematite was not
detected. Hematite was, however, detected after heat treatment at
1050 C as well as increased intensity of the anorthite reflections.
In Fig. 12, SEM observation of the Fe30 glaze after heat treatment at 9001100 C showed both rod shaped and multi-shaped

659

groups of crystals. The rod shaped crystal was identified as anorthite from the XRD and EDS results. The multi-shaped crystals
gradually changed from franklinite to lithium zinc ferrite, which
was mixed with hematite when the heat treatment temperature
increased.
Lo et al.19 reported that in the glass ceramics powder the
crystallization temperature of anorthite was about 880 C. In the
present work, the DTA curve for the Fe30 glaze (Fig. 10) showed
an exothermic peak at 880 C, which should be correlated to
the crystallization peak of anorthite. In line with XRD results
after heat treatment, the exothermic peak from the DTA curve
at 920 C should correspond with the franklinite to lithium zinc
ferrite phase change. When the iron oxide content was high, this
transformation occurred at high temperature.
In glaze containing a high iron oxide content, it was noted
that, after crystallization of franklinite, some excessive Fe3+ was
found as hematite or in lithium zinc ferrite. When the temperature increased, Fe2 O3 may partially dissolve in the melt and an
endothermic peak may show at about 1000 C.2 The resultant
exothermic peak at 1040 C was assumed to be due to the over
saturated Fe3+ ion leading to the crystallization of hematite. This
peak was also observed after the heat treatment at 1050 C.
A crystalline glaze at low temperature was achieved due to
the co-effects of lithium, zinc and iron oxides additions into the
glazes. The suitable composition and flux content lower the melting temperature to 1100 C and the soaking process stimulated
crystallization of zinc and iron oxides. Li+ partially substituted
for Zn2+ in the ZnFe matrix of franklinite (ZnFe2 O4 ) generating an attractive sparkling effect from the lithium zinc ferrite
(Lix Zn12x Fe2+x O4 where x = 0.050.20) crystalline phase in

Fig. 12. SEM images of the Fe30 glaze after heat treatments.

660

A. Wannagon et al. / Journal of the European Ceramic Society 33 (2013) 653660

the glaze. However, the content of iron oxide should not be too
high, because the excess could crystallize as hematite or dissolve
into the glaze when the temperature increased. The latter accelerates the crystallization of anorthite but prohibits the formation
of lithium zinc ferrite.
Romero et al.9 indicated that the iron content has a great
influence on the oxidation state of iron oxide in diopside
glassceramic glazes. Li et al.20 reported that iron oxide acted
as a flux and phase-separation accelerator during firing. The
phase separation induced the crystallization clusters of anorthite surrounded by iron oxide precipitation. The iron oxide
crystal growth developed well due to the severe Al2 O3 depletion and the high iron-content in the inter-crystal phase of
anorthite.
The increase of Al4+ ion replacement for Si4+ in SiO4 tetrahedral feldspar structure might allow more Ca2+ instead of K+
or Na+ to be substituted into the void spaces to maintain the
charge balance. Anorthite was found earlier at low temperature in the high content Fe2 O3 glaze. The phase separation
generated anorthite crystalline clusters rich in alumina content,
subsequent to the formation of hematite crystalline phase at
the cluster edges. This mechanism prohibited the crystallization
of lithium zinc ferrite phase resulting in a decreased sparkling
effect. The heat treatment stimulated Fe2 O3 dissolution into the
glaze and generated a deeper red color as well as promoting
some change from franklinite to lithium zinc ferrite crystalline
phases.
4. Conclusion
The crystalline effect from the glaze containing Li2 O,
ZnO and Fe2 O3 can be obtained at a temperature below
1100 C. Using XRD and SEM/EDS the crystalline phases
were determined as franklinite (ZnFe2 O4 ), lithium zinc ferrite
(Lix Zn12x Fe2+x O4 where x = 0.050.20), hematite (-Fe2 O3 )
and anorthite (CaAl2 Si2 O8 ). The most preferable effect was
the sparkling metallic effect of the lithium zinc ferrite phase
obtained from the glaze containing 10 wt.% of iron oxide (Fe10).
The thermal analysis by STA and heat treatment indicated that
the crystallization temperature of lithium zinc ferrite phase was
860920 C, which increased when the iron oxide content in the
glaze was increased. High content of iron oxide (2030 wt.%)
and its precipitation as hematite and dissolution in the glaze prohibited the formation of lithium zinc ferrite but accelerated the
crystallization of anorthite.

References
1. Knowles KM, Freeman FSHB. Microscopy and microanalysis of crystalline
glazes. J Microsc 2004;215:25770.
2. Levitskii A. Mechanism of phase formation in aventurine glaze. Steklo
Keram 2001;6:2932.
3. Dakhai S, Orlova LA, Mikhailenko NYu. Types and compositions of crystalline glazes (a review). Glass Ceram 1999;56:17780.
4. Dvornichenko IN, Matsenko SV. Production of iron-containing crystalline
glazes. Steklo Keram 2000;2:289.
5. Karasu B, Caki M, Turan S. The development and characterization of zinc
crystal glazes used for Amakusa-like soft porcelains. J Eur Ceram Soc
2000;20:222531.
6. Rudkovskaya NV, Mikhalilenko NYu. Decorative zinc-containing crystalline glazes for ornamental ceramics (a review). Glass Ceram
2001;58:38790.
7. Jamaludin AR, Kasim SR, Ahmad ZA. Effects of different gloss firing temperature on the crystallization of zinc-based crystal glaze. Adv Mater Res
2011;173:448.
8. Jamaludin AR, Kasim SR, Ahmad ZA. The effect of CaCO3 addition on the
crystallization behavior of ZnO crystal glaze fired at different gloss firing
and crystallization temperatures. Sci Sintering 2010;42:34555.
9. Romero M, Rincon JM, Acosta A. Effect of iron oxide content on
the crystallization of a diopside glassceramic glaze. J Eur Ceram Soc
2002;22:88390.
10. Brawer SA, White WB. Structure and crystallization behavior of
Li2 OFe2 O3 SiO2 glasses. J Mater Sci 1978;13:190720.
11. Salman SM, Salama SN. Pyroxene solid solutions crystallized from
CaOMgO (Li2 O, Fe2 O3 )SiO2 glasses. Ceram Int 1986;12:2218.
12. 12.Salman SM, Darwish H, Mahdy EA. The influence of Al2 O3 , MgO and
ZnO on the crystallization characteristics and properties of lithium calcium
silicate glasses and glassceramics. Mater Chem Phys 2008;112:94553.
13. Yekta BE, Alizadeh PA, Rezazadeh L. Synthesis of glass ceramic glazes in
the ZnOAl2 O3 SiO2 ZrO2 system. J Eur Ceram Soc 2007;27:23115.
14. El-Shobaky GA, Abdalla FHA, Fouad Zikry AA. Effect of lithium oxide
doping on zinc ferrite formation. Thermochim Acta 1996;289:819.
15. Zhou ZH, Xue JM, Chan HSO, Wang J. Nanocomposites of ZnFe2 O4
in silica: synthesis, magnetic and optical properties. Mater Chem Phys
2002;75:1815.
16. Wang Z, Schiferl D, Zhao Y, ONeill CHSt. High pressure Raman
spectroscopy of spinel-type ferrite ZnFe2 O4 . J Phys Chem Solids
2003;64:251723.
17. Faria deDLA, Lopes FN. Heated goethite and natural hematite: can Raman
spectroscopy be used to differentiate them? Vib Spectrosc 2007;45:11721.
18. Matson DW, Sharma SK, Philpotts JA. Raman spectra of some tectosilicates
and of glasses along the orthoclaseanorthite and nephelineanorthite joins.
Am Mineral 1986;71:694704.
19. Lo CL, Duh JG, Chiou BS. Low temperature sintering and crystallization behavior of low loss anorthite-based glassceramics. J Mater Sci
2003;68:6938.
20. Li W, Luo H, Jianan L, Jiazhi L, Guo J. Studies on the microstructure of the
black-glazed bowl sherds excavated from the Jian kiln site of ancient China.
Ceram Int 2008;34:147380.