Comment on “Structural, dielectric, and magnetic characteristics of Bi(Ni_0.25Ti_0.25Fe_0.50)O₃ ceramics” [J. Mater. Sci.: Mater. Electron. 27, 1209 (2016)]; “Structural and electrical characteristics of (Co, Ti)-modified BiFeO₃” [J. Mater. Sci.: Mater. Electron. 27, 7115 (2016)]; “Structural, electrical, and magnetic characteristics of Ni/Ti-modified BiFeO₃ lead-free multiferroic material” [J. Mater. Sci.: Mater. Electron. 28, 6673 (2017)]

Abstract

This comment on the above three papers argues that the crystals should be of sillenite type and the correct formula should be Bi₂₅FeO₃₉:MTi (M = Co or Ni) instead of BiFeO₃. Due to the doping, a new type of unit cell, previously unknown, was proposed for such crystals. There is also about 15% of a secondary phase, BiFeO₃, in the sample. This part of the sample should be responsible for the magnetic properties recorded.

1 Results

The papers [1,2,3] present the results of studies on the different physical properties of BiFeO₃ ceramics modified by two different dopants. It is clearly seen that all the studied samples have nearly the same diffraction patterns with a characteristic set of diffraction peaks. Thus, it is obvious that the lattice parameters should be similar for all crystals. However, this is not a case. Table 1 presents all data which are totally different. As a result, it is not possible to accept such different unit cells. Moreover, none of the phases are consistent with those of perovskite structure contrary to the statement made by authors in [3]. Thus, it is necessary to find the source of the evident errors made in the interpretation of the diffraction patterns. Moreover, it is necessary to find the common unit cell for the studied crystals.

Table 1 Lattice parameters of modified BiFeO₃ crystals. Standard deviations are omitted for clarity

The authors of the commented papers supposed that they have synthesized the perovskite BiFeO₃ compound doped by some supplementary ions. However, such a crystal should have the diffraction pattern as presented in Fig. 1 which does not agree with the patterns published in the commented papers. Thus, it is necessary to find an alternative structure, taking into account that the diffraction pattern is a kind of fingerprint of a given crystal.

Fig. 1

The diffraction pattern of BiFeO₃ crystal according to the data from ICSD#29921 (R3c, No. 161, a = 5.5785(2) Å, c = 13.8696(5) Å, [4]) corrected to a ≈ 5.49 and c ≈ 13.3 Å to show the similar 2Θ position of some diffraction peaks in the published pattern

The most simple diffraction pattern was presented in [2]—see below in Fig. 2. It will be a subject of the analysis shown below.

Fig. 2

The diffraction pattern from the paper on Bi(Co_1/4Ti_1/4Fe_1/2)O₃ [2]. For clarity of comparison the original indexes hkl are removed

Fortunately, there is another crystal which has a diffraction pattern similar to that observed by authors of the commented papers. This is a sillenite Bi₂₅FeO₃₉ [5]. Thus, the material in the commented papers should be Bi₂₅FeO₃₉ crystals instead of assumed BiFeO₃. Note, that this compound frequently appears as an impurity phase when bismuth ferrite BiFeO₃ is synthesized [6]. However, in the case of commented papers, the BiFeO₃ phase exists only as an impurity [7].

A simple question arises: is it possible to index the diffraction patterns using the unit cell of this sillenite-type structure? First of all we must verify if the samples have only one phase. Then, it should be verified if the other diffraction peaks can be attributed to the main crystal which may be not strictly cubic. Such deviation from cubic symmetry may result in the appearance of new diffraction peaks as well as changes in the intensity of some peaks. Moreover, the intentional doping may also destroy the basic, parent structure. Unfortunately, the lack of raw diffraction data excludes a more detailed analysis of structure. Thus, we try to estimate the possible solutions of structure using the published figure. Only in the case of crystal from [3] we can use the data furnished by authors (Table 1 in the commented paper).

Many sillenites with different compositions have been synthesised so far. Therefore, in the case of commented papers the general formula can be used as Bi₂₅FeO₃₉:MTi (M = Co or Ni). The crystal structure of Bi₂₅FeO₃₉ is described in cubic symmetry with space group I23 (No. 197) and lattice parameters a = b = c = 10.191 Å (ICSD#257493) [5, 8,9,10,11].

To verify the correctness of supposed existing of sillenite phase in the studied samples, the patterns calculated on the basis of ICSD data on corresponding crystals are presented in Figs. 1 and 3. To facilitate the direct comparison of all patterns, the 2Θ range is the same (20° < 2Θ < 70°).

Fig. 3

The diffraction pattern calculated according to the data from ICSD#257493 for pure Bi₂₅FeO₃₉ crystal in the cubic phase, space group I23 (No. 197), and lattice parameter assumed as a = 10.07 Å

Simple comparison of the patterns from Figs. 2 and 3 indicates that there are supplementary peaks at 2θ ≈ 23.1 and 47.3°, and we observe a lack of diffraction peaks at 2θ ≈ 43.5, 48.9, and 53.9°. These supplementary peaks correspond well to those from BiFeO₃ crystal (Fig. 1). Thus, it is clear that the studied sample is bi-phasic, i.e. contains Bi₂₅FeO₃₉ and BiFeO₃. Note that the experimental pattern is a simple sum of the patterns corresponding to both compounds. One can assume that there is of about 15% of the parasite-phase BiFeO₃. Thus, it seems this part of the sample is responsible for the magnetic properties recorded. The reason for the lack of some diffraction peaks (but present in [1] and [3]) is not clear.

The diffraction patterns from papers [1, 3] are slightly more complicated but preserve the same main set of peaks characteristic for pure sillenite.

The attempt to index the main part of diffraction patterns was successful. There exists one common unit cell for all three crystals. The simple estimations of lattice parameters are presented in Table 2. The introduction of two different ions, Co and Ti or Ni and Ti, results in the deformation of the crystal structure and a creation of a new unit cell. The unit cell of prototype cubic phase, Bi₂₅FeO₃₉, calculated in the monoclinic system is presented for comparison. It is not known how the Co and Ni ions are distributed between both components of the sample studied.

Table 2 Lattice parameters of modified Bi₂₅FeO₃₉ crystals doped by given amount of Co/Ni and Ti ions. Standard deviations are omitted for clarity

Note, that the data from Table 2 concern the basic unit cell similar to that of well-known sillenite phase, but expressed in the monoclinic system which is fully equivalent to the cubic phase. Unfortunately, there was not possible to index the patterns in the cubic system. The main feature of this new unit cell is that the a and b axes are oriented along the face diagonal of parent cubic cell, thus a and b are equal to about √2a_cubic. It is interesting that the monoclinic axis is not the same as preserved c-axis. As I know, such unit cell was not observed until now. The preliminary estimation of the alleged space group for this phase gives two possible symmetries: C1c1 (No. 9) or C12/c1 (No. 15).

It is also clear why the sillenite phase appears in all studied samples instead of assumed simple BiFeO₃. The starting compounds were annealed at the temperature corresponding to the creation of γ-phase of Bi₂O₃, the sillenite-type phase (Fig. 4). Thus, the final compound was based on this phase, i.e. sillenite-one.

Fig. 4

The diffraction pattern calculated according to the data from ICSD#2376 for pure γ-Bi₂O₃ crystal in the cubic phase, space group I23 (No. 197), and lattice parameter a = 10.268(1) Å

Similar comments on the analysis of the data on other modified BiFeO₃ crystals can be found in other journals [12,13,14].