2.2 X-Ray Diffraction
23
then ground and heated for 5 h in air or a flow of oxygen, and either quenched, regular
cooled or slow cooled. The resulting samples had masses of approximately 1–2 g.
This method will be referred to as the “one-pot" synthesis and was used to determine
the impact of the high surface area to volume ratio in the combinatorial samples.
To ensure that the results obtained both by the combinatorial method and the onepot synthesis are significant for other synthesis routes, bulk samples were also made
using a tank reactor as described in Ref. [58]. These samples were made by Aaron
Rowe and his contributions to this project will be pointed out explicitly throughout
this thesis. A continuously stirred tank reactor was used to make mixed manganese–
nickel hydroxide precursors. The metal precursor solutions were made using
NiSO 4 ·6 H 2 O and MnSO 4 ·H 2 O (both 99 % from Aldrich Chemical Co) in such a way
that a variety of nickel to manganese ratios were obtained. The tank reactor was used
to mix NH 4 OH with the metal sulfate solutions at a constant pH of 10.3 throughout
the 10 h reaction under a flow of nitrogen. The precursor was then rinsed with distilled water and dried at 140
◦ C before being mixed with varying amounts of Li 2 CO 3
to yield the desired stoichiometry. These mixtures were heated to 900
◦ C for 10 h in a
box furnace and then either quenched to room temperature between two copper plates
or cooled at a slower controlled rate of 5
◦ C/min (which is comparable to the regular
cooling rate used for the combinatorial samples and will be referred to as such).
2.2 X-Ray Diffraction
2.2.1 High Throughput X-Ray Diffraction (XRD)
of Combinatorial Samples
Due to the large number of combinatorial samples, a Bruker D8 Discover X-Ray
system was used to characterize the vast majority of the combinatorial samples. All
X-ray diffractometers used during this project use Cu–Kα radiation. The Bruker
system has a collimated 0.5 mm wide beam, a Göbel mirror and an area detector.
Each scan was made up of three frames with a 30 % overlap and each frame is
obtained by counting for 300 s. Figure 2.4 shows typical results after stitching the
frames together as well as the outcome of integrating along the arcs. This gave
a scattering angle range of 15–70
◦ and required approximately 15 min per sample.
The Bruker D8 diffractometer is equipped with a sample translation stage that allows
motion in the x, y and z directions such that the scattering patterns from samples
in the combinatorial arrays were automatically measured in sequence. As described
previously, the samples were first transferred onto a silicon (100) single crystal which
served as a zero-background holder during the XRD scans as long as the Bragg
condition for the Si(400) peak was avoided. Although the peak positions determined
using the Bruker diffractometer were extremely accurate, the scattered intensities
were affected by the stitching of frames and the integration along arcs. As such,
Rietveld refinement generally failed for scans produced in this way. The next section
describes the diffractometers used when Rietveld refinement was essential.
23
then ground and heated for 5 h in air or a flow of oxygen, and either quenched, regular
cooled or slow cooled. The resulting samples had masses of approximately 1–2 g.
This method will be referred to as the “one-pot" synthesis and was used to determine
the impact of the high surface area to volume ratio in the combinatorial samples.
To ensure that the results obtained both by the combinatorial method and the onepot synthesis are significant for other synthesis routes, bulk samples were also made
using a tank reactor as described in Ref. [58]. These samples were made by Aaron
Rowe and his contributions to this project will be pointed out explicitly throughout
this thesis. A continuously stirred tank reactor was used to make mixed manganese–
nickel hydroxide precursors. The metal precursor solutions were made using
NiSO 4 ·6 H 2 O and MnSO 4 ·H 2 O (both 99 % from Aldrich Chemical Co) in such a way
that a variety of nickel to manganese ratios were obtained. The tank reactor was used
to mix NH 4 OH with the metal sulfate solutions at a constant pH of 10.3 throughout
the 10 h reaction under a flow of nitrogen. The precursor was then rinsed with distilled water and dried at 140
◦ C before being mixed with varying amounts of Li 2 CO 3
to yield the desired stoichiometry. These mixtures were heated to 900
◦ C for 10 h in a
box furnace and then either quenched to room temperature between two copper plates
or cooled at a slower controlled rate of 5
◦ C/min (which is comparable to the regular
cooling rate used for the combinatorial samples and will be referred to as such).
2.2 X-Ray Diffraction
2.2.1 High Throughput X-Ray Diffraction (XRD)
of Combinatorial Samples
Due to the large number of combinatorial samples, a Bruker D8 Discover X-Ray
system was used to characterize the vast majority of the combinatorial samples. All
X-ray diffractometers used during this project use Cu–Kα radiation. The Bruker
system has a collimated 0.5 mm wide beam, a Göbel mirror and an area detector.
Each scan was made up of three frames with a 30 % overlap and each frame is
obtained by counting for 300 s. Figure 2.4 shows typical results after stitching the
frames together as well as the outcome of integrating along the arcs. This gave
a scattering angle range of 15–70
◦ and required approximately 15 min per sample.
The Bruker D8 diffractometer is equipped with a sample translation stage that allows
motion in the x, y and z directions such that the scattering patterns from samples
in the combinatorial arrays were automatically measured in sequence. As described
previously, the samples were first transferred onto a silicon (100) single crystal which
served as a zero-background holder during the XRD scans as long as the Bragg
condition for the Si(400) peak was avoided. Although the peak positions determined
using the Bruker diffractometer were extremely accurate, the scattered intensities
were affected by the stitching of frames and the integration along arcs. As such,
Rietveld refinement generally failed for scans produced in this way. The next section
describes the diffractometers used when Rietveld refinement was essential.