Characterization of Mn-Doped PIN-PMN-PT Single Crystal Grown by Continuous-Feeding Bridgman Method

: Mn-doped Pb(In 1/2 Nb 1/2 )O 3 -Pb(Mg 1/3 Nb 2/3 )O 3 -PbTiO 3 (Mn:PIN-PMN-PT) single crystals are attractive piezoelectric materials owing to their high mechanical quality factor. However, the single crystal boules grown by the conventional Bridgman method show compositional variation along the growth direction. In particular, the Mn content exhibits large variation due to its severe segregation. To improve the compositional uniformity, we applied the continuous-feeding Bridgman method to the growth of a Mn:PIN-PMN-PT single crystal boule. Then, the composition and property distributions of the boule along the growth direction were evaluated. The results showed that excellent composition and property uniformity were carried out over 80mm in boule length. The ranges of the electromechanical coupling coefﬁcient ( k 33 ) and the piezoelectric coefﬁcient ( d 33 ) were 0.931–0.934 and 1352–1517 pC/N, respectively. The ranges of the mechanical quality factor ( Q m 31 ) and the depolarization ﬁeld ( E d ) were 417–535 and 785–859 V/mm, respectively. The Q m 31 and the E d values were higher than those of the non-doped PIN-PMN-PT single crystals. The continuous-feeding Bridgman method is therefore an effective technique for improving the uniformity of the Mn content. As a result, the Mn:PIN-PMN-PT single crystal grown by the continuous-feeding Bridgman method possesses excellent property uniformity with characteristics suitable for high power piezoelectric applications.

Most of the single crystal boules have been grown by the Bridgman method.However, since the unidirectional solidification cannot avoid compositional segregation, the composition varies within a boule [4][5][6].In particular, Mn content within Mn:PIN-PMN-PT boules severely changes along the growth direction in the elements [7].Therefore, the Mn:PIN-PMN-PT boules exhibit wide variations of properties, such as the piezoelectric constant (d 33 ) and Qm, which are undesirable from the viewpoint of the industrial applications.
The continuous-feeding Bridgman method is one of the sophisticated crystal growth techniques to overcome the compositional segregation issue.The method controls the melt composition by feeding additional materials into the melt during the crystal growth and substantially improves the composition uniformity on PMN-PT and PIN-PMN-PT boules compared with the conventional Bridgman method [8][9][10][11].
In this paper, the growth of a Mn:PIN-PMN-PT single crystal boule was conducted by the continuous-feeding Bridgman method to achieve a stable Mn content within the boule.
The characterization of the single crystal was carried out compared with a non-doped PIN-PMN-PT single crystal.

Crystal Growth
The perovskite ceramics were synthesized from Pb 3 O 4 , MgO, Nb 2 O 5 , TiO 2 , In 2 O 3 , and MnO powder with 3N-4N purity by the columbite precursor procedure, which synthesizes InNbO 4 and MgNb 2 O 6 as precursors.The mixing ratios for the compositions of the fed ceramics are shown in Table 1. Figure 1 shows the schematic furnace configuration for the continuous-feeding Bridgman method [8][9][10].The ceramics were loaded in a platinum crucible with an inner diameter of 80 mm on the support tube, and a <110> seed crystal was set at the bottom of the crucible.The ceramics were melted by the heaters, which surround the crucible at a temperature more than 1300 • C.Then, the crystal growth started after soaking for 16 h.The crucible was lowered at an appropriate speed (less than 1.0 mm/h) with the levitation mechanism, and at the same time, the ceramics were continuously fed into the crucible at a feeding rate with the feeding mechanism.After the continuous-feeding growth of 80 mm in boule length, the remaining melt was solidified without feeding.The boule was grown along the <110> direction, which was the same orientation as the seed crystal.Then, the boule was cooled down to room temperature in the furnace.Finally, the crucible was stripped and the as-grown single crystal boule was obtained.In addition, a non-doped PIN-PMN-PT boule was grown by the same growth method to evaluate the effect of the Mn doping on the properties.

Characterization
The {001} oriented plates with 0.5 mm in thickness were cut at several locations of the grown boules along the growth direction for evaluation.The composition of the plates was measured with an X-ray fluorescence spectrometer (Rigaku ZSX Primus IV).Gold and nichrome layers were formed on both large surfaces of the plates by sputtering as electrodes.The plates were poled under a direct current field of 1200 V/mm.With an impedance analyzer (HP4192A), the free and clamped dielectric properties were measured at 1 kHz and 6.4 MHz in frequency, respectively.The electromechanical coupling coefficient (k 33 ) was calculated from the free dielectric constant (ε 33 T ) and the clamped dielectric constant (ε 33 S ) [12].The piezoelectric coefficient (d 33 ) was measured with a piezo d 33 meter (ZJ-3D).The mechanical quality factor (Qm 31 ) was calculated by the 3dB down method [13] from the impedance curve measured near the resonance frequency with an impedance gain phase analyzer (HP4294A).The polarization hysteresis loop was taken at 10 Hz in frequency with a ferroelectric characteristics evaluation system (Toyo Corporation FCE-3) to determine the depolarization field (E d ) and the internal bias (Ei).

Results and Discussion
Figure 2 shows the Mn:PIN-PMN-PT single crystal boule grown by the continuousfeeding Bridgman method.The boule with the diameter of 80 mm and the length of 145 mm was successfully grown.The boule was black-hued, which was obviously different from the non-doped PIN-PMN-PT single crystal boule [9]. Figure 3 shows the MnO, In 2 O 3 , Nb 2 O 5 , MgO, and TiO 2 content distributions along the growth direction for the boule shown in Figure 1.All distributions exhibited very small variation in the position ranging from 0 to 80 mm where the continuous-feeding Bridgman growth was operated.The MnO, In 2 O 3 , Nb 2 O 5 , MgO, and TiO 2 content were 0.22-0.24wt%, 6.55-6.66wt%, 17.73-17.86wt%, 1.08-1.15wt% and 7.74-7.90wt%, respectively.In contrast, obvious composition changes were found in the position over 80 mm.It should be noted that this boule region was grown without feeding.Only the In 2 O 3 content distribution showed no large variation over the entire boule length, since the In 2 O 3 is not significantly affected by the segregation behavior [7].The results prove that the continuous-feeding Bridgman method was able to improve the uniformity of the MnO content as well as other content for the Mn:PIN-PMN-PT single crystal boules.
Figure 4 shows the piezoelectric properties of the Mn:PIN-PMN-PT boule along the growth direction, compared with the non-doped PIN-PMN-PT boule.The stable properties were achieved in the position ranging from 0 to 80 mm.However, the properties drastically changed in the position over 80 mm.The region ranging from 0 to 80 mm in length had stable composition as shown in Figure 3   The ranges in the electromechanical coupling coefficient k 33 and the piezoelectric coefficient d 33 in the stable property region of the Mn:PIN-PMN-PT boule were 0.931-0.934and 1352-1517 pC/N, respectively.The ranges were slightly lower than those of the nondoped PIN-PMN-PT boule.Therefore, the Mn doping made a negative contribution to piezoelectric performance.Figure 5 shows the ferroelectric hysteresis loops of the Mn:PIN-PMN-PT crystal and the non-doped PIN-PMN-PT crystal.Both crystals show almost the same remnant polarization (~25 C/m 2 ).The Mn:PIN-PMN-PT crystal exhibited an anisotropic loop and an offset along the x-axis, which indicated the existence of internal bias (E i ) [13].The E i was around 90 V/mm.A depolarization field (E d ) is an electric field at which the polarization is zero in the positive field as indicated in Figure 4.The positive field is the reverse polarity for the DC poling.The E d of the Mn:PIN-PMN-PT crystal and the non-doped PIN-PMN-PT crystal were 785 V/mm and 610 V/mm, respectively.The difference between their E d was 175 V/mm, which was almost twice the internal bias.Therefore, the Mn:PIN-PMN-PT crystal had a higher E d than the non-doped PIN-PMN-PT crystal by the internal bias (E i × 2). Figure 6 shows the distributions of the E d and the Ei along the growth direction for the Mn:PIN-PMN-PT boule.The E d and the E i ranged from 785 to 859 V/mm (+/−5% in variation) and from 77 to 92 V/mm (+/−9% in variation), respectively.The excellent uniformity of MnO content shown in Figure 3a contributed to the steady E d and E i within the boule, since the Ei level was associated with the MnO content [14].It was demonstrated that the Mn:PIN-PMN-PT boule grown by the continuous-feeding Bridgman method possessed stable electric field stability.Figure 7 shows the distributions of the Qm 31 along the growth direction for the Mn:PIN-PMN-PT boule.The maximum Qm 31 and the minimum Qm 31 were 535 and 417, respectively.The variation was controlled to +/−12% due to the small composition variation.In addition, the Qm 31 slightly increased along the growth direction.The MnO content also slightly increased along the growth direction, as shown in Figure 3a.Therefore, we think the MnO content strongly relates to the Qm 31 , and the control of the MnO content is essential for stable mechanical quality factors.Table 2 summarizes the typical properties of the Mn:PIN-PMN-PT single crystal and the non-doped PIN-PMN-PT single crystal.Compared with the non-doped PIN-PMN-PT crystal, the Mn:PIN-PMN-PT crystal exhibited low permittivity ε 33 T /ε 0 , the same dielectric loss (tan δ), almost the same k 33 , low d 33 , high E d , and significantly higher mechanical quality factor Qm 31 .Zheng et al. reported that the polarization rotation restricted by the internal bias enhances the mechanical quality factor [15].Therefore, we concluded that the internal bias shown in Figure 4 contributed to increasing the Qm 31 for the Mn:PIN-PMN-PT crystal.The Mn:PIN-PMN-PT crystal is more suitable for high power applications due to its high mechanical quality factor and high depolarization field compared to the non-doped PIN-PMN-PT crystal.

Figure 1 .
Figure 1.Schematic configuration of growth system for continuous-feeding Bridgman method.
Figure3shows the MnO, In 2 O 3 , Nb 2 O 5 , MgO, and TiO 2 content distributions along the growth direction for the boule shown in Figure1.All distributions exhibited very small variation in the position ranging from 0 to 80 mm where the continuous-feeding Bridgman growth was operated.The MnO, In 2 O 3 , Nb 2 O 5 , MgO, and TiO 2 content were 0.22-0.24wt%, 6.55-6.66wt%, 17.73-17.86wt%, 1.08-1.15wt% and 7.74-7.90wt%, respectively.In contrast, obvious composition changes were found in the position over 80 mm.It should be noted that this boule region was grown without feeding.Only the In 2 O 3 content distribution showed no large variation over the entire boule length, since the In 2 O 3 is not significantly affected by the segregation behavior[7].The results prove that the continuous-feeding Bridgman method was able to improve the uniformity of the MnO content as well as other content for the Mn:PIN-PMN-PT single crystal boules.Figure4shows the piezoelectric properties of the Mn:PIN-PMN-PT boule along the growth direction, compared with the non-doped PIN-PMN-PT boule.The stable properties were achieved in the position ranging from 0 to 80 mm.However, the properties drastically changed in the position over 80 mm.The region ranging from 0 to 80 mm in length had stable composition as shown in Figure3due to the continuous feeding.The property variations in the Mn:PIN-PMN-PT boule were small and almost the same as those of the non-doped PIN-PMN-PT boule.The results indicate that the continuous-feeding Bridgman method can realize Mn:PIN-PMN-PT boules with fewer piezoelectric property variations.On the other hand, the properties trended along the growth direction were different between the Mn:PIN-PMN-PT and the non-doped PIN-PMN-PT boules.The Mn:PIN-PMN-PT boule exhibited the largest k 33 and d 33 at around 40 mm in length.In contrast, the non-doped PIN-PMN-PT boule showed the minimum k 33 and d 33 at the same position.This difference was inexplicable only by the composition.The other factors such as domain configuration and residual stress may affect the properties and the contribution to the

Figure 4 .
Figure 4. Distributions of (a) electromechanical coupling coefficients k 33 and (b) piezoelectric constants d 33 along the growth direction for Mn-doped and non-doped PIN-PMN-PT boules grown by the continuous-feeding Bridgman method.

Figure 6 .
Figure 6.Distributions of E d and E i along the growth direction for the Mn:PIN-PMN-PT boule.

Figure 7 .
Figure 7. Distributions of Qm 31 along the growth direction for the Mn:PIN-PMN-PT boule.

Table 1 .
Mixing ratios of oxides used for fed ceramics.

Table 2 .
Typical properties for Mn-doped and non-doped PIN-PMN-PT single crystals.