Experimental Observation of Possible Pressure-Induced Phase Transformation in GdAlO3 Perovskite Using In Situ X-ray Diffraction
, Vadym Drozd 1, Shanece Esdaille 1, Jiuhua Chen 1,*, Surendra Saxena 1, Xue Liang 1 and Leonid Vasylechko 2Round 1
Reviewer 1 Report
This is a very well written paper reporting newly-measured powder XRD data on GdAlO3 to 23 GPa. The publication is significant because a phase transition in GdAlO3 from orthorhombic to tetragonal symmetry is observed at ca. 16 GPa.
I have a couple of comments / questions that should be addressed in a revision before final acceptance:
The experiments used a methanol-ethanol (4:1) mixture as the pressure-transmitting medium. It is well known that methanol-ethanol freezes near 10 GPa and creates a non-hydrostatic environment. This might affect and even induce the phase transition as stress fields can affect the structural distortions in perovskites (see Zhao et al., J. Appl. Crystallography, 2010; Zhao et al., J. Physic: Cond. Matt., 2011). The authors should address this point and include pertinent references.
The high-pressure phase is identified as being tetragonal belonging to the I4/mcm space group. While the refinement of the powder XRD data appears to be consistent with I4/mcm, the peaks are broad and there might be subtle distortions not resolved with the powder data. This needs to be confirmed by single crystal XRD.
Is the transition reversible? If so, is there hysteresis?
Author Response
We appreciate the review's comments which help us to improve the manuscript greatly.
Here is our responses to each of the comments.
1. "The experiments used a methanol-ethanol (4:1) mixture as the pressure-transmitting medium. It is well known that methanol-ethanol freezes near 10 GPa and creates a non-hydrostatic environment. This might affect and even induce the phase transition as stress fields can affect the structural distortions in perovskites (see Zhao et al., J. Appl. Crystallography, 2010; Zhao et al., J. Physic: Cond. Matt., 2011). The authors should address this point and include pertinent references. "
Response: This is very valuable comment. we highlight this fact and cite the references in the revised manuscript as follows.
Line 138-146: We realize that the methanol-ethanol (4:1) pressure medium solidifies at pressure above 10 GPa [45] creating non-hydrostatic pressure environment and therefore the sample may experience significant stresses at 21 GPa. Zhao et al [46,47] were able to quantify the stress field using single-crystal x-ray diffraction and indicate that such a non-hydrostatic pressure environment in a DAC may affect and even induce the phase transition as stress fields can influence the structural distortions in perovskites. A more recent experimental study [48] demonstrates that applying non-hydrostatic pressure may lower the transition pressure in a DAC.
2. "The high-pressure phase is identified as being tetragonal belonging to the I4/mcm space group. While the refinement of the powder XRD data appears to be consistent with I4/mcm, the peaks are broad and there might be subtle distortions not resolved with the powder data. This needs to be confirmed by single crystal XRD. "
Response: This is absolutely true. We added a line to stress this in the manuscript.
Line 151-153: To further confirm the symmetry of the high pressure phase, higher resolution structure characterization, e.g., in-situ single crystal x-ray diffraction, is needed.
3. "Is the transition reversible? If so, is there hysteresis?"
Response: Very important issue that was not mentioned in the original manuscript. Yes, it is reversible as the ambient pressure phase reappears when the pressure is release. We added such a statement as below. Unfortunately, we did not collect the diffraction pattern with fine steps so we cannot specify the hysteresis.
Line 119-121: this transition is reversible as all these disappeared peaks reappear when the pressure is released back to ambient.
Reviewer 2 Report
please find my comments in the attached PDF file
Comments for author File: 
Comments.pdf
Author Response
We appreciate the reviewer's comments greatly. The manuscript is improved significantly in response to these insightful comments. Here are our responses to each of the comments.
- "The phase transition is subtle, I am wondering that the ambient-pressure phase can still do the Rietveld refinement of the XRD collected at 21 GPa. The author should show the failure of the Rietveld refinement of the XRD at 21 GPa with ambient pressure phase, to compare with the successful Rietveld refinement of the XRD at 21 GPa with the I4/mcm phase. That would be more convinced."
Response: This is a very important comment which led a significant revision of the manuscript. Upon the reviewer's suggestion, we actually examined three possible structure models, Pbnm, I4/mcm and R-3c, according to the space group-subgroup theory. It turns out all three models can fit the XRD data reasonably well (with Rw less than 3%). We then performed DFT calculation, and found out R-3c structure is favorable at high pressure in terms of total energy. In the revised manuscript, we include the refinement results of all three models as well as the DFT result. The major reversions are as follows.
Line 121-153: From symmetry relationship point of view, there are three higher symmetry phases for perovskite in the pathway from its Pbnm orthorhombic to cubic polymorphs, i.e., P4/mbm, I4/mcm and R-3c [16]. However, differences in symmetry operations require that the (111) and (021) peaks in the Pbnm orthorhombic phase disappear in the I4/mcm tetragonal phase (i.e., hkl: h+k+l = 2n and 0kl: k, l = 2n) and the R-3c rhombohedral phase while they are allowed in the P4/mbm tetragonal phase. Since the (111) and (021) peaks in the Pbnm orthorhombic phase completely disappeared when sample pressure increased to 21 GPa (Figure 1), the P4/mbm tetragonal phase was not considered. We, therefore, examined three possible structural models using the diffraction pattern at 21 GPa, i.e., Pbnm (the ambient pressure phase), R-3c (the high-pressure phase observed in SmAlO3 [20] and I4/mcm (the structure of CeAlO3). As shown in Figure 2 and Table 1, the differences in structure refinements using the three different space groups are not significant enough to conclusively determine the favorable structure for the high-pressure phase. We then calculated the total energy of the three structure models as a function of pressure using DFT (Figure 3). The calculation results indicate that Pbnm is favorable at low pressures whereas R-3c is favorable at high pressures although the calculated transition pressure (75 GPa) is much higher than the experimental pressure where the disappearance of diffraction peaks was observed. We realize that the methanol-ethanol (4:1) pressure medium solidifies at pressure above 10 GPa [45] creating non-hydrostatic pressure environment and therefore the sample may experience significant stresses at 21 GPa. Zhao et al [46,47] were able to quantify the stress field using single-crystal x-ray diffraction and indicate that such a non-hydrostatic pressure environment in a DAC may affect and even induce the phase transition as stress fields can influence the structural distortions in perovskites. A more recent experimental study [48] demonstrates that applying non-hydrostatic pressure may lower the transition pressure in a DAC. Taking this stress influence and possibly large uncertainty in determining transition pressure using DFT calculations into consideration, we propose that the theoretically predicted Pbnm to R-3c phase transition may occur at a pressure between 16.2 GPa and 21 GPa promoted by non-hydrostatic pressure. As shown in Table 1, the volume per formula unit of the R-3c structure is nearly identical to the Pbnm structure within the experimental uncertainty. To further confirm the symmetry of the high pressure phase, higher resolution structure characterization, e.g., in-situ single crystal x-ray diffraction, is needed.
2. In figure 2, indicate the units of the two axis.
Response: the units are included in the revised figures.
3. In figure 5, the author claimed that there is a 0.5% jump of the unit-cell volume at 21GPa, but it is based on the fitting of the data below 16 GPa. Moreover, the fitting line is within the error bar of the two data above 21 GPa. The author should change the statement.
Response: Thanks for this comments. In the revised figure for the new high pressure structure, the volume discontinuity is nearly unrecognizable. The statement has been changed accordingly.
Line 258-260: the volumetric discontinuity is so small that it is unrecognizable with the experimental uncertainty of this study.
4. In section discussion, the author should explain EXAFS .
Response: "extended X-ray absorption fine structure" is added in Line 437.
5. The linear relationship in figure 6 is poor, if the author really want to show the negative relationship between the bulk modulus of RalO3 and Ionic radius of R3+, the author should add more data in figure 6.
Response: When we consider the data of LaAlO3, PrAlO3, NdAlO3, SmAlO3 and EuAlO3, a good negative relationship is clear. As Figure 6 also include the scattered GdAlO3 data which are not a part of the data demonstrating the relationship, it looks a bit confusion. To make the relationship clear to the readers, we differentiate the data used to define the relationship with open symbols and left the scattered GdAlO3 data as solid symbol in the revised figure. The relationship is better visualized by the open symbols now.
