Quantitative analysis of zinc vaporization from manganese zinc ferrites

Article abstract of DOI:10.1111/j.1151-2916.2003.tb03372.x

Zinc vaporization of Mn-Zn ferrites was quantitatively characterized in terms of oxygen partial pressure P_O2, temperature, grain size and sample geometry. The amount of zinc loss was measured as a function of time at various temperatures by a thermogravimetric method. The weight loss due to irreversible zinc vaporization showed a linear behavior with time and increased exponentially with temperature. The observed weight loss due to zinc evaporation at 1100°C was small, whereas a significant weight change was detected at 1200°C. The weight loss was even greater in a reducing atmosphere (P_O2 = 5 × 10^-5). Below 1300°C, the diffusion of elemental zinc was sufficiently fast to compensate the zinc loss at the interface region, resulting in a linear dependence on time. At temperatures ≥1400°C, the weight change no longer followed the linear dependence and showed a rather parabolic behavior with a concave upward slope. The core shape and the gas flow around ferrite cores were important factors that affected the rate of zinc vaporization, but not the grain size.

Full text of DOI:10.1111/j.1151-2916.2003.tb03372.x

J. Am. Ceram. Soc., 85 [5] 765–68 (2003)
journal
Quantitative Analysis of Zinc Vaporization from
Manganese Zinc Ferrites
Jung Ju Suh and Young Ho Han*
Department of Material Science and Engineering, Sungkyunkwan University, Suwon 440–746, Korea
Zinc vaporization of Mn-Zn ferrites was quantitatively char-
acterized in terms of oxygen partial pressure PO2, temperature,
grain size and sample geometry. The amount of zinc loss was
measured as a function of time at various temperatures by a
thermogravimetric method. The weight loss due to irreversible
zinc vaporization showed a linear behavior with time and
increased exponentially with temperature. The observed
weight loss due to zinc evaporation at 1100°C was small,
whereas a significant weight change was detected at 1200°C.
The weight loss was even greater in a reducing atmosphere
phase enhances grain growth and development of a duplex
7
structure. The initial permeability decreases with increasing zinc
8
loss, whereas tan ␦/␮ and ferrous iron concentration increase. It
o
was also reported that the stacking condition of ferrite cores for
sintering as well as the setter under ferrites have crucial effects on
9
–12
zinc loss and on electromagnetic properties.
Although numerous problems due to zinc loss are recognized,
little work has been done on quantitative analysis of zinc losses
during sintering. Few experimental results have been on EDAX
8
–11
and EPMA studies of the subsurface region of ferrite cores.
؊
5
(PO2 ؍ 5 ؋ 10 ). Below 1300°C, the diffusion of elemental
This study will provide the in situ weight change at high temper-
ature and quantitatively analyze the zinc loss. Zinc loss behavior
will be also discussed in terms of sample geometry and oxygen
activity.
zinc was sufficiently fast to compensate the zinc loss at the
interface region, resulting in a linear dependence on time. At
temperatures >1400°C, the weight change no longer followed
the linear dependence and showed a rather parabolic behavior
with a concave upward slope. The core shape and the gas flow
around ferrite cores were important factors that affected the
rate of zinc vaporization, but not the grain size.
II. Experimental Procedure
Mn-Zn ferrites of a composition Mn0.72Zn0.22Fe2.06O4 were
prepared by a conventional ceramic processing technique. The
ingredient raw materials were mixed in a planetary mill, calcined
at 950°C for 3 h, and then pulverized to obtain a mean particle size
of ϳ1.2 ␮m. The granulated powder was formed into various
shapes as shown in Table I. A thermobalance (Model D101, Cahn,
Madison, WI) was used to measure zinc evaporation from Mn-Zn
ferrites. The sample was hung with Pt wire in a vertical furnace
and then heated to the measurement temperatures with the heating
rate of 250°C/h. A mixture of air and nitrogen gas (P ϭ 5 ϫ
I. Introduction
HE zinc ion preferentially occupies the tetrahedrally coordi-
T
nated A site of the spinel structure in Mn-Zn ferrites; the other
cations occupy either A sites or octahedrally coordinated B sites.
Nonmagnetic zinc ions consequently decrease the magnetic can-
cellation between A and B sites, leading to an increase in the net
O
2
Ϫ5
10 atm (5 Pa)) was used to obtain the desired oxygen activity.
An overall gas flow of 400 cm /min was used to maintain constant
1
3
magnetic moment of Mn-Zn ferrites. Thus, the addition of ZnO is
an essential way to obtain the desirable high permeability of
Mn-Zn ferrites; 18–23mol% of ZnO for high permeability mate-
rials and 8–13mol% for low-loss materials.
PZn and PO2 because the oxygen activity might change the
concentration of the gas phase on the solid surface. The thermobal-
ance coupled with a vertical tube furnace (4.5 cm i.d. and 60 cm
length) was designed for sample weights up to 100 g and is
sensitive to detect weights as small as 1 ␮g. In fact the balance
sensitivity was limited, at high temperatures, to ϳ50 ␮g due to
convection currents. This was sufficient to detect the weight
change of full size ferrite cores at the elevated temperature range.
2
Zinc vaporization has been recognized as troublesome during
sintering of Mn-Zn ferrites, because ZnO is easily reduced and
3
4
evaporates. Slick and Basseches reported that zinc vaporization
is significant at low oxygen activities under high temperature. The
reduction reaction and mass action expression of the zinc vapor-
5
ization can be written as
1
ZnO ϭ Zn͑g͒ ϩ O ͑g͒
(1)
(2)
III. Results and Discussion
2
2
Figure 1 shows the in situ weight change of cylindrically
pressed compacts of ZnO powder as a function of heating time
under air (PO2 ϭ 0.21) and nitrogen (PO2 ϭ 5 ϫ 10 ) at various
K
PZn
ϭ
ͱ
P
Ϫ5
O2
PZn and PO2 are the zinc vapor pressure and the oxygen partial
pressure, respectively. Consequently, zinc vaporization decreases
the Zn content from the original ferrite formula, resulting in
adverse effects on microstructures and electromagnetic properties.
Table I. Weight, Surface Area, and Specific
Surface Area for Specimens of
Different Geometries
3
ϩ
The concentration of ferric ions (Fe ) would be decreased to
6
offset the zinc vaporization. It was reported that the zinc vapor
Weight
(g)
Surface area
Specific surface area
2
2
Shape
(cm )
(cm /g)
Toroid
15
8
27.8
15.1
18.5
8.67
1.85
1.88
2.31
2.89
H. Anderson—contributing editor
8
3
Manuscript No. 187277. Received December 6, 2001; approved December 12,
Cylinder
40
10
32.7
7.89
0.82
0.79
2
002.
Member, American Ceramic Society.
*
7
65

7
66
Journal of the American Ceramic Society—Suh and Han
Vol. 85, No. 5
Fig. 1. Weight losses from pure ZnO compacts as a function of time for
various temperatures.
temperatures. The weight decreases linearly with time and the
slope of weight change versus time depends on the process
temperature and oxygen activity. A small weight loss was ob-
served at 1100°C in nitrogen, whereas there was a significant
change in the slope at 1300°C when the zinc evaporation rate
was ϳ 0.1 wt% per hour. These results are in good agreement with
the zinc vapor pressure expressed in Eq. (2), which is inversely
proportional to the square root of oxygen partial pressure and
proportional to the equilibrium constant, i.e., increasing exponen-
Ϫ6
tially with temperature. The zinc vapor pressures, 9.66 ϫ 10 atm
Ϫ3
(
966 kPa) and 1.63 ϫ 10 atm (163 Pa) were calculated at 1100°
and 1300°C, respectively. The linear dependence of weight change
on time implies that the vaporization rate was independent of
diffusion processes such as zinc migration to the sample surface,
and the zinc vapor pressure on the surface of specimens was
maintained at a constant value. This result is acceptable because
3
the vertical process tube (ϳ950 cm ) was flushed with enough gas
Fig. 2. Weight losses from pure Mn-Zn ferrites as a function of time for
various temperatures.
3
(450 cm /min) to ensure the constant ambient pressure condition.
Mn-Zn ferrites are nonstoichiometric oxides consisting of
transition-metal ions and their ionic state is easily changed. The
oxidation state of these specimens could be affected by tempera-
ture and oxygen partial pressure (P ). Thus, the weight change
whereas the zinc loss is irreversible. Thus, the net weight loss due
to the zinc evaporation could be separated from the total weight
change. The reversible process does not develop any permanent
weight change, but the zinc evaporation would result in permanent
weight losses.
O
2
due to oxidation–reduction during sintering should be taken into
1
3
account. The related equation can be written as:
2
3
3ϩ
1
3ϩ
3
1
2Ϫ
O
2
2ϩ
1
2ϩ
1
Fe ϩ Mn ϩ V ϩ O ϭ Fe ϩ Mn ϩ O
C
2
3
8
2
3
3
4
Figure 3 shows a sequential weight change of a ferrite core at
two different levels of oxygen activity, air and nitrogen (P ϭ
O
2
(3)
Ϫ5
5
ϫ 10 atm) at 1100° and 1200°C as a function of sintering time.
VC and OO represent the cation vacancy and the oxygen ion on
the normal oxygen lattice site, respectively. The forward direction
represents the reduction reaction; half mole of oxygen releases
from the spinel crystal together with the corresponding disappear-
ance of cation vacancies. Thus, the forward reaction results in
weight loss, the reverse reaction leading to weight gain.
The specimen attained an equilibrium state within a reasonable
period of time at 1100°C. After switching to air, the specimen
gained the weight and recovered its original state with a slight
weight loss. However, the specimen did not recover its previous
state at 1200°C, leaving behind a permanent weight change of
ϳ0.05%, which is equal to the linear weight loss followed by the
abrupt change due to the reduction reaction. This confirms that the
compositional deviation of Mn-Zn ferrites after heating is attrib-
utable to zinc evaporation.
Figure 4 shows a series of weight changes under a periodic
atmosphere control: alternating oxygen activities with a fixed
pattern of 9.5 h reduction in nitrogen and 6 h oxidation in air at
1200°C. The repeated oxidation processes made the ferrite cores
completely oxidized and recovered to their previous state except
for the zinc evaporation during the reduction step. It should be
noted that the maximum points after oxidation and the minimum
points after reduction lie on the straight lines with the same slope.
This result supports the assumption that the oxidation–reduction
reaction does not cause any permanent weight change and the
alternating PO2 switch does not change the rate of Zn-evaporation.
However, as the sintering time increases at 1200°C, the density
and microstructure of ferrite cores would change because of the
Figure 2 shows the weight change of ferrite cores after switch-
ing from air to nitrogen at 1100°–1300°C. Unlike the weight
change of ZnO as shown in Fig. 1, no weight change was observed
in air even at 1300°C, and drastic weight changes occurred just
after switching to nitrogen, followed by a linear weight change
with time. The abrupt weight change is presumably due to the
reduction reaction as described in Eq. (3). It should be noted that
the abrupt weight change (%) at 1100°C is greater than at 1200°
and 1300°C. This is in good agreement with the previous result
that there is an abrupt change of slope in the equilibrium ferrite
1
4
weight change with temperature as the temperature is lowered.
At 1100°C, in a nitrogen atmosphere, the specimen gets to an
equilibrium state within a reasonable period of time (Ͻ3 h).
However, equilibrium states were not attained under nitrogen at
1
200° and 1300°C. Such a continuous weight change is attributed
to zinc loss. The oxidation–reduction reaction is reversible,

May 2003
Quantitative Analysis of Zinc Vaporization from Manganese Zinc Ferrites
767
Fig. 3. Weight change of a ferrite core under periodic atmosphere control
at the 1100° and 1200°C isotherms.
progress of sintering. Figures 5(a) and (b) show the microstruc-
tures of the specimens sintered for 1 h and 35 h, respectively, at
1
200°C,. An obvious difference can be seen between them in
densification amounts with significant grain growth. The specimen
shown in Fig. 5(b) was sintered under the alternating oxygen
activity for 35 h as scheduled in Fig. 4. It is an interesting result
that microstructural developments such as grain growth and
densification do not affect the zinc loss kinetics as well as the
oxidation–reduction reaction rate, the maximum and minimum
points lying on the straight line.
Fig. 5. SEM micrographs showing microstructures of fractured surfaces
of Mn-Zn ferrites sintered at 1200°C for (a) 1 and (b) 35 h.
The weight change due to zinc losses shows a linear time
dependence at 1200°C as shown in Fig. 2. This result implies that
the diffusion of elemental zinc from bulk to surface is sufficiently
fast to compensate the zinc loss at the specimen surface, similar to
water evaporation in swimming pools. At 1400°C the weight
change no longer follows a linear dependence as shown in Fig. 6.
The rate of weight change decreases with time, showing a rather
parabolic behavior, and the abrupt weight change, which normally
occurs just after switching to a nitrogen atmosphere, is not
observed. The weight change due to the reduction reaction on
switching from an air to a nitrogen atmosphere seems negligibly
small compared with the zinc loss. The disappearance of the abrupt
weight change is attributed to the trivial change in the oxidation
degree as reported in the phase equilibrium diagram of Mn-Zn
ferrites. The fast evaporation rate depreciates the ZnO concen-
tration at the surface region, lowering the amount of vaporization.
The evaporation rate depends on the ZnO content; the more the
1
4
Fig. 4. Weight change of a ferrite core under periodic atmosphere control
at the 1200°C isotherm.
Fig. 6. Weight loss from Mn-Zn ferrites with time at 1400°C.

7
68
Journal of the American Ceramic Society—Suh and Han
Vol. 85, No. 5
vaporization is an irreversible reaction, it was possible to separate
the net amount of zinc weight loss by subtracting the weight
change due to oxidation–reduction from the total weight change.
Thus, the irreversible process results in permanent weight losses.
At 1100°C a negligibly small weight loss was observed with zinc
evaporation, whereas at 1200°C a significant weight change due to
zinc loss was detected. The weight loss stemming from zinc
vaporization increased exponentially with temperature. Below
1
300°C, the weight loss was a linear function of time, indicating
that the diffusion of elemental zinc to the surface of specimens is
fast enough to compensate the zinc loss at the interface region.
However, at 1400°C the weight change no longer followed the
linear dependence and showed a rather parabolic behavior with a
concave upward. It was confirmed that the rate of zinc vaporiza-
tion is strongly influenced by the core shape and the gas flow
around ferrite cores but is not affected by the grain size.
Fig. 7. Weight loss per hour (%) as a function of specific surface area for
toroidal and cylindrical ferrite cores.
References
1
S. Chikazumi, Physics of Ferromagnetism, 2nd Ed.; Ch. 9. Oxford University
Press, New York, 1997.
2
A. Goldman, Modern Ferrite Technology; Ch. 11. Van Norstand Reinhold, New
ZnO on the surface area the higher the rate of weight loss. It thus
leads to a parabolic behavior of zinc vaporization because the bulk
diffusion of ZnO to the surface area controls the evaporation
kinetics. However, weight loss due to zinc evaporation, obtained
from the total weight change less that due to the reduction reaction,
still has an important meaning from a practical point of view. A
processing condition as severe as heating at 1400°C is not common
in real sintering processes.
Figure 7 shows the weight change per hour as a function of
specific surface area of toroidal and cylindrical ferrite cores. As
summarized in Table I, the surface area of specimens per gram
depends on sample geometry. It is interesting that the cylindrical
specimen shows larger weight changes than those estimated from
the extrapolation of toroidal specimens. The vapor pressure de-
pends on the surface curvature. Toroidal specimens have a concave
surface (inner radius) as well as a convex surface (outer radius),
whereas cylindrical specimens have only a convex surface. In
addition, the gas flow would be stagnant at the inner side of the
toroid cores. The bare specimens showed five times greater
evaporation rate than cores put in an alumina crucible. These
results indicate that the core shape and gas flow around ferrite
cores are important factors affecting zinc vaporization.
York, 1992.
3
A. Beer and J. Schwarz, “Neue Ergenisse Uber den Einfluss der Sinterbedingun-
gen Auf,” IEEE. Trans. Mag., 2 [3] 470–74 (1966).
4
P. I. Slick and H. Basseches, “Thermogravimetric Study of the Solid–Gas
Interaction of a Mn-Zn Ferrite and the Effect on its Magnetic Properties,” IEEE.
Trans. Mag., 2 [3] 603–606 (1966).
5
G. Schnitt and P. Kleinert, “Studies on Material Transport during the Formation
of Zinc and Nickel Ferrite from the Oxides. II.,” Z. Anorg. Allg. Chem., 398, 41–53
(
1973).
6
D. W. Johnson, P. K. Gallerher, M. F. Yan, and H. Schreiber, “Vacancy Contents
and Phase Equilibria of Ti-Substituted Mn-Zn Ferrite,” J. Solid State Chem., 30,
2
99–310 (1979).
7
P. Sainamthip and V. R. W. Amarakoon, “Role of Zinc Volatilization on the
Microstructure Development of Manganese Zinc Ferrite,” J. Am. Ceram. Soc., 71 [8]
6
44–48 (1988).
8
H. J. Hellicar and A. Sicignano, “Dynamic Role of Zinc Oxide in the Sintering of
Manganese Zinc Ferrites,” Am. Ceram. Soc. Bull., 61 [4] 502–505 (1982).
9
S. H. Chen, S. C. Chang, I. N. Lin, M. J. Tung, and W. B. Shu, “Effect of Sintering
Parameters on Zn Loss for Preparation of High Permeability Mn-Zn Ferrites,” IEEE.
Trans. Mag., 28 [5] 2436–38 (1992).
10
Y. Okazaki, Y. Kitano, and T. Narutani, “Interaction between Mn-Zn Ferrites and
2 3
Al O Refractory during Sintering”; pp. 313–16 in Ferrites, Proceedings of the 6th
International Conference on Ferrites. Japan Society of Powder and Powder Metal-
lurgy, Tokyo, Japan, 1992.
1
1
S. H. Chen, S. C. Chang, and I. N. Lin, “Zinc Loss of Mn-Zn Ferrite Sintered
under Alumina Plate,” IEEE. Trans. Mag., 32 [5] 4857–59 (1996).
12
M. J. Tsay, M. J. Tung, C. J. Chen, and T. Y. Liu, “The Manufacture of
High-Permeability Mn-Zn Ferrites by Atmospherical Protect,” J. Phys. IV France, 7,
C1.71–C1.72 (1997).
1
3
R. Morineau, “Structual Defects and Oxidation–Reduction Equilibrium in Mn-Zn
Ferrites,” Phys. Staus. Soidi A, 38, 559–568 (1976).
IV. Conclusion
1
4
P. I. Slick, “A Thermogravimetric Study of the Equilibrium Relations between a
Mn-Zn Ferrite and an O -N Atmosphere”; pp. 81–83 in Ferrites, Proceedings of the
Zinc vaporization from Mn-Zn ferrites was quantitatively ana-
lyzed by a thermogravimetric technique. The amount of zinc loss
was determined by a periodic change of atmosphere. As the zinc
2
2
1
st International Conference on Ferrites. Edited by Y. Hoshino, S. Iida, and M.
Sugimoto. University of Tokyo Press, Tokyo, Japan, 1970.
Ⅺ