Page 1
*
Undergraduate Student – Department of Materials Science and Engineering – Lehigh University
Tel.: 610-597-8007
E-mail:[email protected]
Sintering Behavior of 1% and 10% concentrations of
Zinc, Copper, and Boron Doped Titania
Under Conventional Sintering
George J. Ferko V*
Department of Materials Science and Engineering, Lehigh University, 5 East Packer Avenue, Bethlehem, PA, 18015, United States
April 24, 2009
Abstract
The role of dopants in the sintering behavior of ceramics has been a major theme in the studying of ceramics for the last 50 years. The
industrial popularity of titania has prompted this study of its sintering behavior under 1% and 10% concentrations of zinc (Zn), copper
(Cu), and boron (B) dopant, conventionally sintered at temperatures of 1000oC and 1200
oC for sintering times of 0, 2, and 6 hours.
The study has provided insight into the formation of second phase, the growth of abnormal grains, and the effect of the level of
doping, sintering temperature, and sintering time on the densification and grain size in the final samples. The Cu doped samples were
found to have second phase present in all samples. The Zn doped samples had second phase in all samples sintered at 1200oC. The B
doped samples were all found to have abnormal elongated grains. Longer sintering time always resulted in larger grains. Higher
doping levels always lead to larger grains except in the Zn doped samples where the formation of second phase inhibited grain growth.
Higher sintering temperature always lead to larger grains in the samples tested. The doped samples typically exhibited a higher
densification than the undoped samples. Densification was inhibited in the B doped samples due to the elongated shape of the grains.
Densification increased with an increase in sintering time in the Zn doped samples that were sintered at 1000oC.
Introduction (background and justification)
(1)
Titania has many applications through a broad range of
industries. Titania has the greatest use in products which
involve white pigments. The high refractive indices and high
reflectance of anatase and rutile make them perfect for making
pigments white and bright. White pigment applications for
TiO2 exist in the coatings, paper, ink, paints, plastics, rubbers,
ceramics, fibers, ultraviolet light protection, food, and
cosmetics industries. TiO2 is considered to be thermally
stable, nontoxic, noncorrosive, and noncombustible making it
irreplaceable in the many sectors of industry to which it is
applied. Titania offers a high amount of opacity meaning the
color of other aspects of a product cannot be seen through it.
TiO2 is also used as a photocatalyst because of the ability of
its band gaps to absorb ultraviolet radiation through excitation
of valance electrons. Photocatalytic applications include
cleaning of contaminated water and self cleaning windows and
tiles [1][2]. TiO2 is of particular importance for the modeling
of ceramics because of the broad range of industry to which it
is applied. The study of the effects of different dopants on the
sintering behavior of TiO2 will allow for better production of
the products made with TiO2. The modeling of the sintering
behavior of TiO2 anatase will also provide insight into the
sintering behavior of other ceramics with the tetragonal crystal
structure [3].
(2)
It is important to control the microstructure (grain size
and shape) and the densification (percent porosity) in sintering
because the final microstructure and porosity will have strong
effects on the properties of the ceramic. The mechanical,
thermal, electrical, magnetic, and optical properties are all
strongly affected by final densification and microstructure.
Being able to control the grain growth and the densification
allows for the creation of advanced ceramics with properties
better than those of conventional metals and ceramics [4][5].
(3)
Mehdi Mazaheri, et al, have done significant research into
the various sintering methods of commercially pure titania and
their effect on grain size and densification in 2008 [4]. Shen
Dillion, et al, have done research and created theories about
the role of grain boundary complexions into grain boundary
mobility and the formation of abnormal grains in 2007 and
2008 [6][7]. Chak Chan, et al, have done research on the
effects of different calcinations temperature on the
microstructure of sintered ceramics in 1999 [7]. Yu-hong
Page 2
2 George J. Ferko V
Zhang, et al, have done research on the effects of various
metal dopants and the ability to inhibit grain growth in titania
in 2002 [9][10].
(4)
The goal of this experiment is to study the effects of
doping elements (Zn, Cu, and B), doping level, sintering
temperature, and holding time on the densification and grain
growth during the sintering of TiO2 ceramics [11].
Experimental Methods
(1)
3 grams of TiO2 anatase powder from Alfa Aesar with a
starting particle size of about 150 nm were measured on a
Mettler P1200 scale in a Nalgene low density polyethylene
(LDPE) tray and poured into a Nalgene LDPE bottle. 1 mol/L
of Zn(NO3)2 in ethanol is measured in a graduated cylinder
and poured into the LDPE bottle. The amount of dopant
needed is determined by equation 1 seen on page 3. An extra
10 mL of ethanol is poured in the bottle and the solution is
mixed and un-conglomerated ultrasonically in the VWR
B3500A-DTH for 10 minutes. The bottle was labeled and
placed under a Labconco fume hood overnight to evaporate
the ethanol. Some samples that remained wet had their drying
accelerated by putting a light over them. The sample was
poured into a double bag and excess was scrapped out of the
bottle with a clean plastic knife. The agglomerates were
broken by rolling a cylindrical metal weight over the bags.
The sample was put into an alumina crucible and placed in the
Lindberg/Blue M box furnace model # BF51866A-1 where it
was calcined at 600oC for 2 hours with a heating and cooling
rate of 5oC/min. The total time of the calcinations comes out
to six hours; however, the samples were left in the furnace
overnight and removed the next day after which they were
poured back into their plastic bag. 0.8 grams of powder was
poured from the 10%Zn-TiO2 onto the Mettler P1200 scale in
a Nalgene HDPE tray. Agglomerates were broken with a
plastic fork which had been kept in a bag designated for the
10% Zn doped powders. The powder was poured into a tool
steel die and placed into the Carver Laboratory Press model:M
and pressed at very low pressure. The pellet was released and
die was held carefully while the cylindrical ring used to push
out the pellet was put on and the die and ring were put back in
the press. Pressure was reapplied slowly until the pellet
emerged and was placed into a latex glove. The air was
sucked out of the glove using a Welch – Chem Star – Model:
1400N – vacume pump so that isostatic pressure could be
applied. During pressing the glove was tied shut and the
pellets were isostatically pressed at 45000 psi for 3 minutes in
the Fluitron cold isostatic press. The pressure is built up and
released slowly. After pressing the pellet it will be 40-60% of
the theoretic density. Excess oil was washed off the gloves
with hot water and pellets were cut out of each finger in the
glove. The pellet was placed in a Pall sample holding plastic
tray. A layer of TiO2 anatase powder was poured into an
alumina crucible as a base layer under the Zn-TiO2 pellets and
an additional layer of TiO2 anatase powder was poured on top.
The pellets were placed into the Thermo Scientific –
Thermolyne – TSM furnace where they will be held at
1000oC for 2 hours with a heating and cooling rate of 5
oC/min.
The following day the samples were removed from the furnace
and returned to their Pall sample holding plastic trays. An
epoxy mounting container was cleaned with ethanol and
release agent was swabbed onto the mounting surfaces of both
parts of the two part mounting container. Hardener was mixed
with the epoxy resin for 2 minutes until it was clear and then
poured into the mounting mold on top of the sample. The
sample was placed into a vacuum and vacuum was created and
released twice (for 2 cycles). The sample was placed under a
fume hood to allow it to cure for 24 hours. A Buehler Ltd.
Metallurgical Apparatus grinding wheel was used on low
speed for the grinding and polishing of the sample. The
sample was hand grinded and polished using 320, 400, 600,
8µm, and then 3µm SiC paper followed by 1µm diamond
paste. The samples were cleaned with ethanol and then
placed, while in ethanol, into an F350 ultrasonic for 5 minutes
to be cleaned. Vibratory polishing was done with the Buehler
Vibromet I polisher using a 0.15 μm SiO2 colloidal suspension
for 24 hours. Sample was removed and sprayed with ethanol.
The epoxy sample was then heated in the L and L box furnace,
so that fumes would be carried away by the hood. Pellet was
removed from the epoxy while it was still brittle. The pellet
was returned to an alumina crucible for thermal etching. The
Thermo Scientific – Thermolyne – TSM furnace was heated
to 900oC where temperature was held for 30 min with a
heating and cooling rate of 5oC/min. After thermal etching the
sample was removed and placed on SEM mounting stubs
using carbon tape. The samples were coated with Iridium in
the Electron Microscopy Sciences – Model: EMS5575X –
peltier cooled Iridium coating machine for 8 sec. Using the
JEOL 840 SEM with an accelerating voltage of 15kV and a
working distance of 15mm micrographs of the sample were
taken at the top edge, middle, and bottom edge. Magnification
was determined subjectively by considering how well the
images would provide accurate data. Print outs of these
micrographs were used to graphically analyze the sample for
grain size, using equation 5, and percent porosity. The
abnormal or normal grain growth was observed subjectively.
(2)
Page 3
3 George J. Ferko V
Volume of the doping solution is a function of c, x, m,
and M, where c is the molar concentration of the doping
solution, x is the doping level, m is the mass of the TiO2
powder, and M is the TiO2 molar mass. The equation used to
calculate V, the volume of the doping solution is shown below
in equation 1.
(1)
(3)
The doped TiO2 powders must be calcined in order to
remove any volatile elements that may remain from the doping
process in week one. In week one the doping agents are added
to the TiO2 anatase powder in solution to produce a good
distribution of dopant throughout the anatase powder. The
ethanol in the solution is evaporated from the solution and
then calcination takes place according the reactions below,
shown in equations 2, 3, and 4.
(2)
(3)
(4)
As can be seen, the doping elements are not the only thing
initially added to the TiO2 anatase powder, but additional
atoms remain for each dopant atom. The calcination process
removes the additional atoms by raising the temperature of the
sample so that those additional atoms become volatile and
leave the sample. The calcination temperature is chosen by
finding a temperature that is higher than the decomposition
temperature so that the free energy of the system after the
reaction is lower than the free energy before. The
decomposition temperature for the various dopant precursors
is determined by thermal gravitational analysis (TGA). In this
test the sample mixed with dopant is heated while on a scale
until the weight of the sample decreases. The temperature that
is chosen must be as low as possible to prevent the formation
of agglomerates. The temperature must be below the
temperature at which anatase goes through a phase change into
rutile to prevent rutile phase from forming. The temperature
must also be below the temperature at which the crucible
begins to significantly diffuse into the TiO2 [8][12]. The same
temperature must be used for all the samples so that the initial
conditions for sintering are the same for all dopant-TiO2
systems.
(4)
Samples are embedded in TiO2 anatase powder because of
the high energy state that the samples will be in when heated
to their respective temperatures. If the samples were not
embedded in powder particles from the sample would be
excited and released into the air around the sample in the
furnace. The samples are embedded to prevent any diffusion
of the alumina crucible into the titania sample. Also the cost
of maintaining an inert atmosphere in a furnace for sintering is
very high. If the doped TiO2 samples are embedded in titania
powder than a non-inert atmosphere will not have an
opportunity to react with the actual sample. The embedding of
the samples allows them to be sintered without the conditions
inside the furnace, other than temperature, effecting the final
microstructure and densification.
(5)
Samples must be etched after polishing in order to reveal
their grain boundaries so that their microstructure may be
studied. The thermodynamic principle working here is that
when the surface of the sample is excited using high
temperature the highest energy material will diffuse to a lower
energy state outside of the sample. The highest energy
material happens to be on the samples grain boundaries‟, so
heating the sample proves to be a good way to reveal the
microstructure.
(6)
The sample pellet does not need to be embedded in titania
powder during thermal etching. This is so because the
sintering of the sample was performed at a higher temperature
than the thermal etching. This means that the bulk of the
material in the sample energetically prefers to stay in the
sample at temperatures even higher than the thermal etching
temperature. The only material that will leave the sample
during thermal etching is the material on the grain boundaries.
The temperature of thermal etching is not high enough for
diffusion from the crucible into the sample to occur at
significant levels. If the sample were embedded in powder
than some of the powder would remain on the surface of the
sample making valuable imaging in the SEM impossible.
(7)
The equation used to calculate grain size from the SEM
images is shown below in equation 5. The method used is
referred to as the linear interception method. The value
obtained from this two dimensional analysis is slightly skewed
from the actual three dimensional value. The skew of the data
is not taken into consideration because the two dimensional
analysis is just as accurate at comparing relative grain size.
180mm is the length of the measuring line, n is the number of
times the measuring line hit a grain boundary, µm is the length
of the micron bar as it is read from the image, and mm is the
length of the micron bar on the print. The grain size from this
equation will be in units of µm.
(5)
Page 4
4 George J. Ferko V
Results and Discussion
(1)
a. b.
c. d.
Figure 1: a. TiO2 doped with 10% Zn and sintered for 2 hours at 1000oC, b. TiO2 undoped and sintered for 2 hours at 1000
oC, c. TiO2
doped with 1% Zn and sintered for 2 hours at 1000oC, d. TiO2 doped with 10% Zn and sintered for 2 hours at 1200
oC.
(2)
The 10% Zn – TiO2 sample that was sintered for 2 hours
at 1000oC can be seen in figure 1a. Grains in the sample have
grown to a size of 0.88 µm with the peak in grain growth rate
occurring somewhere near 2 hours, as can be seen in figure 2
below on page 5. The sample has the 10th
smallest grain size
of all of the 42 samples tested and the 4th
smallest grain size
out of all of the 12 Zn doped samples. The 10% Zn – TiO2
sample that was sintered for 2 hours at 1000oC is the 17
th most
dense sample over all of the 42 samples tested and the 5th
most
dense over all of the 12 Zn doped samples with a percent
porosity of 5%. Although the porosity data may be inaccurate
do to the small number of measurements taken, it can be
assumed that this sample is somewhere in the middle range as
Page 5
5 George J. Ferko V
far as relative densification goes. For a description of the
effects of temperature, sintering time, and doping levels on
densification refer to the results and discussion section,
subsection (5) on page 6. Original particle size as well as the
presence of agglomerates plays a large role in the densification
process. The compaction of the green pellet also plays a large
role in the final amount of densification. These variables have
not been taken into consideration in this experiment. The
grain growth of the 10% Zn – TiO2 sample that was sintered
for 2 hours at 1000oC was much less than those performed at
higher temperatures and for longer times, figures 4 and 5.
This does fit the grain coarsening theory which states that
higher temperature and longer time result in larger grains. All
of the undoped samples for the sample temperature have a
smaller final grain size indicating that the Zn doping facilitates
grain growth in some way by changing the energy associated
with the grain boundary interface and the diffusivity across the
grain boundaries [4].
(3)
Throughout the lab there was ample opportunity for the
introduction of impurities into the sample. Impurities consist
of oil from students‟ hands, oil from the isostatic press, other
ceramic materials in the lab rooms that were worked in, and
atmospheric gases from the furnaces that did not have inert
gas atmospheres. The rooms that the sample was handled in
were not clean rooms and the hood it was left under to
evaporate ethanol out of the original doping solution was not a
clean hood. The containers that the sample was stored in may
not have been cleaned as well as is possible. Any time the
sample was subject to heating it was not in an inert
atmosphere thus there may have been some reaction with the
atmospheres in the furnaces used for calcination, removal of
the epoxy sample mount, and thermal etching. Accidental
dropping the sample or contamination of equipment used
throughout the lab, such as plastic knives, plastic forks, plastic
bags, crucibles, and sample trays could have led to
contamination of the sample. Accidentally touching the
sample while trying to mount it on the SEM stub or while
moving it between containers may also have caused the
sample to be contaminated. Contamination after sintering is
not as important as contamination before sintering.
Contamination before sintering may have affected the
diffusion properties of the sample. It is likely that some
contaminates would have sat on the grain boundaries causing a
change in grain growth and densification. The contaminates
would act as impurity defects and change the local driving
forces around them. The effect of the contamination has been
ignored for the most part because of high concentrations of
dopant being used in the doping process. Because such a large
concentration of dopant is used little to no effect from
contaminates should be noticeable in the samples. These
effects that are present may have been marginally avoided by
looking for consistency in microstructure across the samples
during the taking of micrographs in the SEM. By avoiding
analysis of abnormal sections of the pellet and taking
micrographs of the microstructures that appeared to be most
common in the sample any abnormal grains caused by
contamination should not have been included in the porosity
and grain size data.
(4)
Figure 2: Plot of the grain size versus the sintering time for TiO2 samples that were undoped, doped with 1% Zn, and doped with 10%
Zn and sintered at 1000oC, as well as samples that were undoped, doped with 1% Zn, and doped with 10% Zn and sintered at 1200
oC.
0
1
2
3
4
5
6
0 2 4 6 8 10 12 14 16 18 20
Sin
terin
g T
ime
(hrs
)
Grain Size (µm)
Grain Size vs. Sintering Time
undoped 1000
undoped 1200
1%Zn 1000
1%Zn 1200
10%Zn 1000
10%Zn 1200
Page 6
6 George J. Ferko V
(5)
The temperature at which sintering takes place seems to
play a large role on densification. All of the Zn doped
samples that have a lower percent porosity then the 10% Zn –
TiO2 sample that was sintered for 2 hours at 1000oC have been
sintered at 1200oC, figures 4 and 5. This indicates that a
greater amount of densification can be seen with temperatures
above 1000oC. None of the undoped samples achieved a
lower percent porosity than the 10% Zn – TiO2 sample that
was sintered for 2 hours at 1000oC. This indicates that the
dopant had some effect as to increase the densification;
however, because of the inaccuracy of the data this effect
cannot be pin-pointed. Among the Zn doped samples those
with a longer sintering time most often have a higher
densification than those with a shorter sintering time. Overall,
the Zn and Cu doped samples have greater densification than
the undoped and B doped samples, the samples with longer
sintering times have greater densification than those with short
sintering times, and the samples with higher sintering
temperatures have greater densification than those with lower
sintering temperatures. This is interesting data because it is
expected that grain and pore coarsening will occur at higher
temperatures and longer sintering times. In the samples tested
this does not occur.
(6)
The doping level will change the position on the phase
diagram that is dealt with during the sintering process. This
change in position does not affect the phase changes that the
Zn-TiO2 goes through. The micrographs of the 10%Zn-TiO2
taken of the samples that were sintered at 1200oC indicate that
a second phase forms in the grain boundaries of the sample. It
is unclear what the composition of the second phase is because
no analysis was done on it, but it is clear that the second phase
forms due to the level of doping and the sintering temperature.
The doping level will affect the concentration of defects that
take occur and thus affect the diffusivity in the sample. A
higher concentration of dopant makes extrinsic defects,
equation 6, become more likely relative to intrinsic defects,
equation 7, shown below. We know that for all practical
ceramics the defect reaction in equation 6 is the more likely to
occur.
(6)
(7)
Increase in the number of defects is known to lead to an
increase in the diffusivity of a ceramic and thus will change
the sintering behavior. From the micrographs of the samples
we can see that increasing the amount of dopant seems to
Figure 3: Phase diagram for the ZnO-TiO2 system showing all
phase transformations above the transformation from anatase
to rutile crystal structure [13].
increase the amount of abnormal grains, figures 4 and 5. The
amount of Zn dopant does have some effect on densification.
The low temperature Zn doped samples decrease in porosity
over time compared to the undoped low temperature samples
which do not change a large amount over time, figure 4. The
high temperature Zn doped samples have been highly
densified previous to the times at which they are observed and
the undoped samples have densified much less, figure 5. The
data here indicates that doping facilitates the densification of
TiO2 in some way. It is likely that this occurs because the
additional defects that occur due to the presence of dopant
effect the diffusion through the lattice and the interfacial
energy. The effect of diffusion through the lattice and
interfacial energy on densification rate can be deduced from
the Herring scaling law seen below in equation 8.
(8)
Grain size increases with the addition of Zn dopant. The grain
growth kinetics that apply to grain size are very similar to
Page 7
7 George J. Ferko V
those that apply to densification. The addition of dopant
causes an increased number of defects and thus changes the
interfacial energy at the grain boundary as well as the
diffusivity across the grain boundary. This concept is shown
in equation 9.
(9)
The concentration of dopant will effect two terms in the
numerator of the equation so we can expect the effect of
concentration of dopant on grain growth and final grain size to
be exponential. In the high temperature 10% Zn doped sample
the grain growth is restricted by the presence of a second
phase. For the high temperature the sample with the lower
concentration of Zn has a larger average grain size than the
sample with a high concentration of Zn. The second phase
lowers the diffusivity that occurs in the perpendicular
direction across the grain boundary and changes the interfacial
energy between the grains causing the change in grain growth.
A significant amount of impingement may have also occurred
to inhibit grain growth. The two temperatures the samples are
sintered at are both in the Zn2TiO4+Rutile phase on the phase
diagram. This should show that the temperature has no effect
on the phase that is present during sintering. This is not the
case. The high temperature 10% Zn doped sample has a
second phase present in the microstructure where lower
temperature sample does not. The increase in temperature will
increase the defect concentration causing similar effects to
grain growth and densification as the increase in dopant
concentration. The equation that shows the effect of
temperature on the extrinsic defect concentration is shown
below in equation 10.
(10)
Equations 8 and 9 indicate that the increase in temperature
should lower densification and grain growth due to the
temperature term in the numerator. The effect that
temperature has on the defect concentration, diffusivity, and
interfacial energy is great and so the grain growth and
densification actually increase with temperature. This can be
seen in the micrographs taken of the samples, figures 4 and 5.
The effect of sintering time on grain size can be found by
taking the integral of equation 9. This new equation can be
seen below in equation 11.
(11)
This equation shows that the grain size relative to the grain
size at a sintering time of zero increases linearly with
increasing sintering time. When the grains are inhibited from
growing due to a second phase or impingement the grain
growth is slowed. Sintering time increases the densification
and can be seen in the micrographs of the low temperature Zn
doped samples. The densification of the Zn doped samples at
high temperature does not seem to vary with sintering time;
however, the densification in the doped samples is much
greater than in the undoped samples. This occurs in the high
temperature samples because the samples have already
densified to a large extent and if there is very little porosity
remaining than the sample cannot continue to densify at a high
rate.
(7)
The undoped samples increase in grain size with
increasing sintering time and increase in densification with
increasing temperature, figures 4 and 5. The increase in
sintering time does not appear to have an effect on
densification. Grain size increases with increasing
temperature. The Cu doped samples, figures 6 and 7, increase
in grain size with increasing sintering time. The effect of
sintering time on densification is unclear because inaccuracy
of the data. The higher sintering temperature increases the
grain size in the Cu doped samples a large amount, 10 to 20
micrometers in most cases. Higher sintering temperature does
not appear to have a significant effect on densification for the
Cu doped samples. The concentration of dopant in the Cu
doped samples has a large effect on the sintering behavior.
Second phase in the high temperature sample appears to be
slowing down the rate of grain growth. This can be deduced
by noting that the difference in grain size between the 10% Cu
doped sample and the 1% Cu doped sample is much less than
the difference in grain size between the 1% Cu doped sample
and the undoped sample. According to equation 11 this is an
indication that the grain boundary mobility is decreasing due
to formation of second phase. Concentration of dopant
doesn‟t have a clear relationship to densification; however, the
Cu doped samples are clearly less porous than the undoped
samples. The B doped samples, figures 8 and 9, increase grain
size with increasing sintering time. The effect of sintering
time on densification in the B doped samples is unclear from
the data. The grain size of the B doped samples increases with
increasing sintering temperature and the effect of sintering
temperature on densification is unclear from the data. The
grain size increases with concentration of dopant and the
effect on densification due to concentration of dopant is
unclear from the data. It should be noted that the abnormal
needle shaped grains that develop for the B doped samples
likely keep them from densifying with a trend connected to
higher temperature, sintering time, or concentration of dopant.
The abnormally shaped grains in the B doped sample occur
Page 8
8 George J. Ferko V
Sintering time: 0hours 2 hours 6 hours
a.
Porosity: 20.80%, Grain size: 0.17µm Porosity: 20.30%, Grain size: 0.32µm Porosity: 20.80%, Grain size: 0.47µm
b.
Porosity: 22.22%, Grain size: 0.54µm Porosity: 16.20%, Grain size: 0.86µm Porosity: 0.07%, Grain size: 0.99µm
c.
Porosity: 24.89%, Grain size: 0.54µm Porosity: 5.00%, Grain size:0.88µm Porosity: 5.89%, Grain size: 1.21µm
Figure 4: a. undoped samples sintered at 1000oC, b. 1%Zn doped samples sintered at 1000
oC, c. 10% Zn doped samples sintered at
1000oC.
Page 9
9 George J. Ferko V
Sintering time: 0hours 2 hours 6 hours
a.
Porosity: 2.56%, Grain size: 3.45µm Porosity: 4.12%, Grain size: 10.5µm Porosity: 3.00%, Grain size: 19.0µm
b.
Porosity: 2.67%, Grain size: 3.49µm Porosity: 0.56%, Grain size: 8.75µm Porosity: 0.94%, Grain size: 12.48µm
Figure 5: a. 1% Zn doped samples sintered at 1200oC, b. 10% Zn doped samples sintered at 1200
oC.
Sintering time: 0hours 2 hours 6 hours
a.
Porosity: 0.94%, Grain size: 6.10µm Porosity: 6.10%, Grain size: 8.30µm Porosity: 1.89%, Grain size: 11.63µm
b.
Porosity: 20.20%, Grain size: 6.83µm Porosity: 3.60%, Grain size: 25.55µm Porosity: 4.67%, Grain size: 28.57µm
Figure 6: a. 1% Cu doped samples sintered at 1000oC, b. 10% Cu doped samples sintered at 1000
oC.
Page 10
10 George J. Ferko V
Sintering time: 0hours 2 hours 6 hours
a.
Porosity: 3.40%, Grain size: 30.00µm Porosity: 0.00%, Grain size: 37.44µm Porosity: 2.17%, Grain size: 41.19µm
b.
Porosity: 8.83%, Grain size: 38.66µm Porosity: 1.00%, Grain size: 41.92µm Porosity: 4.50%, Grain size: 45.45µm
Figure 7: a. 1% Cu doped samples sintered at 1200oC, b. 10% Cu doped samples sintered at 1200
oC.
Sintering time: 0hours 2 hours 6 hours
a.
Porosity: 17.30%, Grain size: 0.80µm Porosity: 18.33%, Grain size: 0.885µm Porosity: 19.60%, Grain size: 0.92µm
b.
Porosity: 14.70%, Grain size: 0.84µm Porosity: 17.70%, Grain size: 0.893µm Porosity: 23.86%, Grain size: 1.07µm
Figure 8: a. 1% B doped samples sintered at 1000oC, b. 10% B doped samples sintered at 1000
oC.
Page 11
11 George J. Ferko V
Sintering time: 0hours 2 hours 6 hours
a.
Porosity: 25.70%, Grain size: 1.02µm Porosity: 33.33%, Grain size: 1.45µm Porosity: 21.78%, Grain size: 1.66µm
b.
Porosity: 15.11%, Grain size: 1.43µm Porosity: 16.30%, Grain size: 1.59µm Porosity: 15.20%, Grain size: 2.08µm
Figure 9: a. 1% B doped samples sintered at 1200oC, b. 10% B doped samples sintered at 1200
oC
due to a grain boundary complexion transition that must occur
at some point in the sintering process. For a given dopant
there is a certain concentration of that dopant necessary to
cause a complexion transition. The likelihood of excess
dopant and transition increases with increasing grain growth.
It is possible that some abnormal grains may form due to
extrinsically caused inhomogeneity, but in the case of the B
doped TiO2 the abnormal grains occur very frequently
throughout the sample. This confirms that the abnormal grain
growth is due to at least two grain boundary complexions
coexisting during grain growth in the sample [6][7].
(8)
The function of dopants in sintering and the mechanics by
which they affect the microstructure of ceramics has been a
long debated issue. The concept of complexion has begun to
solve many discrepancies in the models that predict the
behavior of ceramics with the addition of different
concentrations of dopant. The concentration of dopant has
been found to cause a certain complexion on the grain
boundary which increases grain boundary mobility. The
change in grain boundary mobility due to different doping
concentrations as well as the effect of changing sintering
temperature and sintering time on the grains is the basis for
this lab. The complexion of the grain boundaries not only
explains the sintering behavior and the change in sintering
behavior with sintering time and temperature, but it also
explains the formation of abnormally shaped grains and the
interaction of grains with a second phase. The concept of
complexions can explain why the Zn doped TiO2 grains have
not become abnormal at the contact points with the second
phase. The same explanation can be used for the grains that
are in contact with second phase in the Cu doped sample.
Complexions explain that it is not the saturation of dopant
along the grain boundary that causes abnormality in the grain.
It is the complexion of the grain boundary that will affect its
growth. The grain boundary mobility of all of the samples is
directly related to the growth rate of the grains. This shows
that studying the complexions of Cu, Zn, and B doped TiO2
would help to explain the mechanisms by which their grains
grow [6][7].
Conclusions
This study has provided insight into the formation of second
phase, the growth of abnormal grains, and the effect of the
level of doping, sintering temperature, and sintering time on
the densification and grain size in the final samples. The
following conclusions were made from observing the
micrographs of the doped TiO2 as well as porosity and grain
size data:
1. The Cu doping of titania samples caused second
phase to be present in all samples.
Page 12
12 George J. Ferko V
2. The Zn doping of the samples caused second phase in
all samples sintered at 1200oC.
3. The B doing of samples will cause formation of
abnormal grains.
4. Longer sintering time always leads to larger grains.
5. Higher doping level always leads to larger grains
except in the Zn doped samples where the formation
of second phase inhibits grain growth.
6. Higher sintering temperature always leads to larger
grains in the samples tested.
7. The doped samples will typically exhibit a higher
densification than the undoped samples.
8. Densification will be inhibited in the B doped
samples due to the elongated shape of the grains.
9. Densification will increase with an increase in
sintering time in the Zn doped samples when sintered
at 1000oC.
10. Elongated grains in the B doped samples will be
caused by the coexistence of two or more grain
boundary complexions making one side of the grain
have greater grain boundary mobility than the other.
This occurs due to a transition in grain boundary
complexion during the sintering process.
Future Work
(1)
It is suggested that before future analysis of the Cu, Zn,
and B dopants in TiO2 is done that more data be collected.
Data should be collected for more concentrations of the
dopants by diffusing the dopants through the samples to
develop a concentration gradient of the dopant. The samples
should also be sintered for shorter intervals of time as well as
longer times. The sintering temperatures for the samples
should vary more. Specifically the samples could be sintered
at a higher temperature if alumina crucibles were not used and
zirconia crucibles were used instead. After data is collected
for all of these variables analysis of complexions could be
done in the TEM by milling out sub-100 nm films using the
FIB and taking images of the grain boundaries. By this
method the different types of grain boundary complexion and
the conditions for transitions in grain boundary complexion
could be found for the TiO2 systems. Discovery of the
different types of complexion in the TiO2 systems would
explain the actual mechanics behind the sintering of these
samples. Hot pressing may also be used as the sintering
process to study other effects of the dopants.
(2)
On the atomic level complexion is caused by the different
atomic radii and valance charges between the dopant and the
material to be doped. Using this scenario it can be assumed
that other dopants that will be good for inducing grain growth
will have similar valance charges and atomic radii to those
used in this lab. This is assumed because the dopants used in
this lab caused increased grain growth suggesting that the
complexions created increase grain boundary mobility. To
form similar complexion to that created by Zn, Cu, and B
dopants such as Silver (Ag), Cadmium (Cd), Indium (In),
Gallium (Ga), Aluminum (Al), Silicon (Si), and Carbon (C).
It is very possible that some of these suggested dopants will
not facilitate grain growth at all, however, because the actual
complexions have not been evaluated in this lab no conclusion
can be made about exactly what dopants will enhance grain
growth.
(3)
In three separate sintering papers on Si-doped TiO2 the Si-
base precursor was tetraethyl orthosilicate, Si(OC2H5)4
[9][10][14]. This precursor is mixed with the TiO2 powder in
butanediol, C4H10O2, and calcined. The proper method for
selecting a dopant is to check if the doping will result in the
release of toxins. If that does not occur than the solubility of
the dopant in the solution and material to be doped should be
checked. If only dopants that release toxins are soluble with
the material to be doped than they must be used. Once these
major factors help to narrow down dopants other factors such
as availability and cost may be factored into a decision about
choosing a dopant. The dopant picked above was not arrived
at by this process. This dopant was found in literature on
doping TiO2. The precursor chosen for doping with alumina is
aluminum nitrate, Al(NO3)3. This precursor produces no
toxins and is soluble in water. The precursor can be mixed
with titania in water and dried to produce an even distribution
of dopant. During calcinations no toxins will be released. The
precursor is also readily available from Alfa Aesar.
(4)
This lab was very interesting. The amount of outside
research required and the length of the lab really helped to
reinforce many of the concepts learned about in the class. My
only wish is that the lab was coming to an end at a slightly
more convenient time.
References
[1] Richard Walton, "Titanium oxides", in
AccessScience@McGraw-Hill,
http://www.accessscience.com, DOI 10.1036/1097-
8542.801320.
[2] “Chemistry Sectors: Colorants & Fillers – Titanium
Dioxide Manufacturers Association (TDMA) – TiO2
– Uses and Properties”, cefic, 2009.
Page 13
13 George J. Ferko V
http://www.cefic.be/Templates/shwAssocDetails.asp?
NID=473&HID=25&ID=173.
[3] “Titanium Dioxide”, Wikipedia, 2009, <
http://en.wikipedia.org/wiki/Titanium_dioxide>.
[4] Mehdi Mazaheri, et al, “Sintering of Titania Nanoceramic:
Densification and Grain Growth”, Ceramics
International 35, 2009, pp 685-691.
[5] J. Rubio, et al, “Characterization and Sintering Behavior
of Submicrometer Titantium Dioxide Spherical
Particles Obtained by Gas-Phase Hydrolysis of
Titanium Tetrabutoxide” Journal of Materials
Science 32, 1997, pp 643-652.
[6] Shen J. Dillon, et al, “Complexion: A New Concept for
Kinetic Engineering in Materials Science”, Acta
Materialia 55, 2007, pp. 6208-6218.
[7] Shen J. Dillon, et al, “Demystifying the Role of Sintering
Additives with „Complexion‟”, Journal of the
European Ceramic Society 28, 2008, pp. 1485-1493.
[8] Chak K. Chan, et al, “Effects of Calcination on the
Microstructures and Photocatalytic Properties of
Nanosized Titanium Dioxide Powders Prepared by
Vapor Hydrolysis”, J. Am. Ceram. Soc. 82 (3), 1999,
pp 566-572.
[9] Yuhong Zhang, et al, “Nanotubes in Si-Doped Titanium
Dioxide”, Chem. Commun. 2002, pp 606-607.
[10] Yu-Hong Zhang, et al, “Phase Transformation and Grain
Growth of Doped Nanosized Titania”, Materials
Science and Engineering C 19, 2002, pp 323–326.
[11] ShuaiLei Ma, “Sintering of TiO2 – Spring 2009 -
Objective”, pp. 1.
[12] “Calcination”, Wikipedia, 2009,
<http://en.wikipedia.org/wiki/Calcination>.
[13] J. Yang, et al, “The Phase Stability of Zn2Ti3O8”,
Materials Characterization 37, 1996, pp 153-159.
[14] Jeerapong Watthanaarun, et al, “Titanium (IV) Oxide
Nanofibers by Combined Sol–gel and
Electrospinning Techniques: Preliminary Report on
Effects of Preparation Conditions and Secondary
Metal Dopant”, Science and Technology of
Advanced Materials 6, 2005, pp 240–245.