REPOSITORY > RESULTS

Abstract

The KNN solid solution was prepared from pre-synthesized binary compounds of,
sodium and potassium niobate in a 1/1 molar ratio. The KNG sintering aid was prepared
by homogenizing the pre-reacted binary mixtures of alkaline reagents and germanium
oxide with the Na/Ge = 1/1 and K/Ge = 1/1 molar ratios. The melting point of the KNG
was at about 720°C as determined by the heating stage microscope. The KNG, heated to
1000oC and cooled to room temperature, was amorphous. The wetting angle of the KNG
at 790oC was as low as 8o, therefore we concluded that KNG would be a suitable sintering
aid for KNN.
The KNN and x KNG, x = 0,5, 1, 2, 4 wt.%, denoted as KNN+xKNG, powder
mixtures were homogenized, compacted and sintered at 1000°C, 8 h. The relative density
of pure KNN was 90,0 % theoretical density (TD), and of KNN+0,5KNG 93,7 %TD,
while the densities of KNN with larger contents of KNG were 95,6 – 95,9 %TD. The
microstructure of KNN was characterized by a broad grain size distribution, with the
mean size of 9 μm, while the microstructures of KNN+1KNG, KNN+2KNG and
KNN+4KNG were uniform with about 4 μm large grains. The KNN+0,5KNG exhibited
exaggerated grain growth, which we connected with a too low amount of the liquid phase
and its non-uniform distribution between the KNN grains. The energy dispersive X-ray
spectroscopy (EDXS) in scanning electron microscope confirmed the presence of
germanium-rich phase on KNN grain boundaries and grain junctions.
The KNN and KNN+KNG samples were further sintered at 1100°C, 8 h. The relative
densities of KNN, KNN+0,5KNG and KNN with larger contents of KNG were 93,2
%TD, 94,0 %TD and about 92 %TD, respectively. The microstructures of the KNN,
KNN+0,5KNG, KNN+1KNG and KNN+4KNG, sintered at 1100°C, were similar to the
counterparts sintered at 1000°C, only the mean grain sizes in the KNN+KNG were
slightly larger. The grain size of the KNN+2KNG sample was noticeably larger, about 12
μm. A possible explanation for such strong size increase could a combined effect of the
liquid phase and partial incorporation of germanium ions into the perovskite lattice.
According to X-ray diffraction (XRD), KNN, sintered at 1100°C, is single phase
perovskite with a monoclinic unit cell.
We observed splitting of all perovskite diffractions in the XRD-patterns of the
KNN+KNG samples, sintered at 1100°C, which we could explain with the coexistence of
two perovskite phases. By Rietveld analysis the cell parameters of the phases were
determined. The volume of the »primary« perovskite in all KNN+KNG samples was
close to the volume of the perovskite unit cell of pure KNN, while the volume of the unit
cell of the »secondary« perovskite was gradually decreasing with increasing KNG
content. This could be explained as the consequence of the partial incorporation of Ge4+
on the Nb5+ sites in the perovskite lattice. Such peak splitting has not been observed in the
XRD-patterns of the KNN+KNG, sintered at 1000°C. The presence of germanium in
KNN grains was further confirmed by EDXS analysis in transmission electron
microscope. We found that majority of KNN grains contained up to about 0,5 at.%
germanium, while some rare grains contained a few times larger amount, about 3,5 at.%.
We assume that the solid solubility of germanium in KNN depends on the matrix grain
size, namely, it is larger in smaller grains, and vice versa, according to Gibbs-Thompson
relation.
According to the pore-filling theory the KNN+ KNG should reach almost theoretical
density (99,4 %TD) after sintering at 1000oC, for 8 hours. We decided to model the
sintering of KNN+2KNG in order to explain the discrepancy between the theoretical
value and the experimental value of about 96 %TD. The samples were sintered at 1000oC
for 0, 2 and 8 h. The results showed that the sintering of KNN+KNG could not be well
described by the pore-filling theory. We found that pore growth was strongly promoted
with increasing time, which could be due to Ostwald ripening and/or to presence of a gas
within the pores, presumably as a consequence of evaporation of alkaline species.
The dielectric permittivity, losses, piezoelectric d33 constant and planar coupling
coefficient kp of KNN+KNG, sintered at 1000°C, 2 h, were about 500, 0,04. 110 pC/N -
120 pC/N and 0,3 – 0,4, depending on the KNG content. The obtained dielectric
permittivity and losses are comparable to the reported values for KNN ceramics, sintered
at about 1100 oC with about 95 % TD, while the planar coupling coefficient kp of 0,40 and
even more the piezoelectric d33 constant with the maximum value of 120 pC/N are higher
than the values reported for KNN (ε/ε0 290 – 650, tanδ 0,02 – 0,06, kp about 0,3, d33 about
90 pC/N). Furthermore, the obtained values of the KNN+KNG, sintered at 1000°C, 2 h,
are comparable to the reported values for KNN with various liquid phase sintering aids
(K4CuNb8O23, K5.4Cu1.3Ta10O20, K1.94Zn1.06Ta5.19O15), nevertheless it should be noted that
these materials were sintered between 1070°C and 1120°C.