2.1 Sample Preparation
21
the changes seen in the combinatorial samples also occur in bulk samples is critical
and is dealt with in Chap. 7. The main consequence of having stearic acid present is
that there are contaminants in the samples, though some contaminants are also due
to not rinsing the samples after the co-precipitation reaction, and all were found to
disappear during synthesis. Figure 2.2 (b)–(d) shows the samples at various stages of
heating. After drying at 55
◦ C, the stearic acid can be seen on the substrate, while at
300
◦ C, decomposition of the stearic acid begins. The fact that the substrate appears
clean after heating to 800
◦ C shows that decomposition of the stearic acid does reach
completion. There is no evidence that the contaminants during heating affect the
final products, nor is there any evidence of there being consequences to mixing all
three metals in one step, but again it is important to confirm that the tank reactor
method gives comparable results to those obtained for the combinatorial samples,
and so bulk and combinatorial samples will be compared throughout this thesis.
Chapter 3 shows the results of testing a number of variables in order to optimize
the combinatorial synthesis method to minimize lithium loss. These variables were:
the substrate used (alumina, alumina coated with LiAlO 2 and magnesia), the precipitator used (ammonium bicarbonate or ammonium hydroxide), the atmosphere used
during heating (air or pure oxygen), the amount of solutions dispensed (10 or 20 μL)
and the temperature at which the samples are heated (200, 300, 400 . . . 1000
◦ C). The
sources of all chemicals were the same as those listed above, and ammonium hydroxide was obtained from Fisher while magnesia plates were obtained from Ceramatec.
All substrates were first coated with stearic acid. The best combination obtained in
that study was used throughout the rest of the thesis and proceeds as follows. Each
combinatorial sample, with a mass of approximately 2 mg, was made by dispensing
a total of 20 μL of 1.78 M solutions with the solution-processing robot. The concentrations were measured to within 2 % using atomic absorption as described later
in this chapter. The three solutions were lithium, manganese and nickel nitrates, and
the amounts of each were varied in order to map out the Gibbs triangle. The substrate
used during heating was alumina. After dispensing the nitrates, 23 μL of 2 M ammonium bicarbonate was added, thereby ensuring that it was in excess for all samples.
The samples were then dried overnight at 55
◦ C before being heated for 3 h at 800
◦ C.
Some samples were heated in air in a box furnace while others were heated in a tube
furnace under a flow of at least 30 mL/min of oxygen. Four cooling rates were used.
The first involved turning off the power to the furnace which will be referred to as
regular cooling. At high temperatures, this resulted in an approximate cooling rate of
8
◦ C/min and an overall rate of about 5
◦ C/min. This cooling method is comparable
to that used in the making of commercial electrodes. The second cooling method
was to quench the samples by transferring the alumina plate from the furnace onto a
steel slab as quickly as possible. A Mastercraft infrared temperature sensor was used
to determine that the quenched samples reached 100
◦ C within 1 min after removal
from the furnace. This corresponds to a cooling rate of roughly 10
◦ C/sec. A second
quenching method was used wherein samples were transferred into liquid nitrogen.
This cooling occurs at the order of a few seconds such that a cooling rate of about
100
◦ C/sec was obtained. The final cooling rate, used occasionally, was to cool the
samples at a controlled rate of 1
◦ C/min and will be referred to as slow cooling.
21
the changes seen in the combinatorial samples also occur in bulk samples is critical
and is dealt with in Chap. 7. The main consequence of having stearic acid present is
that there are contaminants in the samples, though some contaminants are also due
to not rinsing the samples after the co-precipitation reaction, and all were found to
disappear during synthesis. Figure 2.2 (b)–(d) shows the samples at various stages of
heating. After drying at 55
◦ C, the stearic acid can be seen on the substrate, while at
300
◦ C, decomposition of the stearic acid begins. The fact that the substrate appears
clean after heating to 800
◦ C shows that decomposition of the stearic acid does reach
completion. There is no evidence that the contaminants during heating affect the
final products, nor is there any evidence of there being consequences to mixing all
three metals in one step, but again it is important to confirm that the tank reactor
method gives comparable results to those obtained for the combinatorial samples,
and so bulk and combinatorial samples will be compared throughout this thesis.
Chapter 3 shows the results of testing a number of variables in order to optimize
the combinatorial synthesis method to minimize lithium loss. These variables were:
the substrate used (alumina, alumina coated with LiAlO 2 and magnesia), the precipitator used (ammonium bicarbonate or ammonium hydroxide), the atmosphere used
during heating (air or pure oxygen), the amount of solutions dispensed (10 or 20 μL)
and the temperature at which the samples are heated (200, 300, 400 . . . 1000
◦ C). The
sources of all chemicals were the same as those listed above, and ammonium hydroxide was obtained from Fisher while magnesia plates were obtained from Ceramatec.
All substrates were first coated with stearic acid. The best combination obtained in
that study was used throughout the rest of the thesis and proceeds as follows. Each
combinatorial sample, with a mass of approximately 2 mg, was made by dispensing
a total of 20 μL of 1.78 M solutions with the solution-processing robot. The concentrations were measured to within 2 % using atomic absorption as described later
in this chapter. The three solutions were lithium, manganese and nickel nitrates, and
the amounts of each were varied in order to map out the Gibbs triangle. The substrate
used during heating was alumina. After dispensing the nitrates, 23 μL of 2 M ammonium bicarbonate was added, thereby ensuring that it was in excess for all samples.
The samples were then dried overnight at 55
◦ C before being heated for 3 h at 800
◦ C.
Some samples were heated in air in a box furnace while others were heated in a tube
furnace under a flow of at least 30 mL/min of oxygen. Four cooling rates were used.
The first involved turning off the power to the furnace which will be referred to as
regular cooling. At high temperatures, this resulted in an approximate cooling rate of
8
◦ C/min and an overall rate of about 5
◦ C/min. This cooling method is comparable
to that used in the making of commercial electrodes. The second cooling method
was to quench the samples by transferring the alumina plate from the furnace onto a
steel slab as quickly as possible. A Mastercraft infrared temperature sensor was used
to determine that the quenched samples reached 100
◦ C within 1 min after removal
from the furnace. This corresponds to a cooling rate of roughly 10
◦ C/sec. A second
quenching method was used wherein samples were transferred into liquid nitrogen.
This cooling occurs at the order of a few seconds such that a cooling rate of about
100
◦ C/sec was obtained. The final cooling rate, used occasionally, was to cool the
samples at a controlled rate of 1
◦ C/min and will be referred to as slow cooling.