Crystal Chemistry and Electronic Properties of the Al-Rich Compounds, Al2Cu, ω-Al7Cu2Fe and θ-Al13Fe4 with Cu Solution
Round 1
Reviewer 1 Report
From the topic, the reviewed manuscript may be suitable for Metals. Nevertheless, the manuscript has more drawbacks than advantages. I analysed in detail the part considering the Al13Fe4 phase and I cannot recommend it for the publication.
- The substitution of Al by copper is very difficult to understand from the poimt of view of chemical bonding, about which the authors are talking at several places. From the crystal chemistry of intermetallic phases, cu should substitute rather iron. the substitution of AL by the d-metals is usually observed for the ternary systems with an elelctropositive component like rare-earth metal. Thus the suggested Al-by-Cu substitution should be proven carefully, which unfortunaately is nor made in the present manuscript.
- The claimed reduction of the total energy by 0.003 eV/atom (ca 0.3 kJ/mol) is not the value, which may decide about the probability of the substitution at this position.
- The authors refer to the experimental work, where the same Al7 and Al9 positions were suggested for Cu sbstitution. First, in this paper, the composition of the experimental samples wer selected a priori on the line toward Cu, which should model the Cu/Al substitution, but this was also not proven. Second, the the results of the crystal structure refinement are given without the information, which would allow to judge the substitution as relevant; no other experimental data which could be considered as pro substitution are presented. Consequently, this publication is not suitable for verification of the calculational results in the reviewed manuscript.
- Furthermore, there are several fragments in the text, which indicate that the auhors do have difficulties with the basics of physical chemistry. Otherwise, they would not write the following statements, which are quite senseless, if not simply wrong: i) 'Due to its small solubility in Al, Fe exists in the form of Fe-containing intermetallic compounds (Fe-IMCs), including the primary θ-Al13Fe4.'; ii) 'this structure is chemically flexible and Fe/Al can be partially replaced to form ternary compounds'; iii) ' Genba and co-workers prepared Al-Fe alloy under Ar protection in a quartz (SiO2) vessel using pure Al and Fe raw materials [26]. They first heated the sample at 850 ºC and then cooled it to 750 ºC which they quenched it into water. They observed many needle-like crystals appeared on the surface of the SiO2 glass tube. Using the X-ray diffraction patterns analysis and electron probe microanalysis, they found Cu at the Fe1, Fe2, Fe3 and Al9, Al12, Al14 sites in θ-Al13Fe4 [26]'; iv) 'We present the results of the stable θ-Al76CuVII2Fe24 phase'; etc.
Author Response
Responses to Reviewer 1
Q: From the topic, the reviewed manuscript may be suitable for Metals. Nevertheless, the manuscript has more drawbacks than advantages. I analysed in detail the part considering the Al13Fe4 phase and I cannot recommend it for the publication.
A: We appreciate the efforts and the scrutiny of Reviewer 1, reading, understanding, and arising constructive critical comments about our work. Here, in this revised version, we tried lifting any misunderstanding, to make our manuscript more transparent and clearer.
Q1: The substitution of Al by copper is very difficult to understand from the point of view of chemical bonding, about which the authors are talking at several places. From the crystal chemistry of intermetallic phases, cu should substitute rather iron. the substitution of AL by the d-metals is usually observed for the ternary systems with an electropositive component like rare-earth metal. Thus, the suggested Al-by-Cu substitution should be proven carefully, which unfortunately is nor made in the present manuscript.
A1: We agree with Reviewer 1 that it is difficult for some readers to understand why Cu substitutes for Al rather than Fe. We have described this in Section 3.1. We showed that the Cu 3d orbitals are fully occupied and positioned lower in energy (~2.3eV below the Fermi level). Thus, they should be better treated as (semi)core electrons and they contribute little to the chemical bonding with neighboring atoms. Correspondingly, Cu behaves differently from Fe which has partially filled 3d orbitals. Chemical bonding occurs mainly between the Cu 4s, 4p electrons and the orbitals of the neighboring atoms. There is a weak interaction between Cu 3d electrons and Fe 3d orbitals (Section 3.2 and Fig. 3). Thus, the energy costs of the Cu substitution for Fe in θ-Al13Fe4 are high (Table 3). The interaction between Cu 4s, 4p and Al 3s, 3p orbitals originates from the Al-by-Cu substitution (Section 3.1).
Q2: The claimed reduction of the total energy by 0.003 eV/atom (ca 0.3 kJ/mol) is not the value, which may decide about the probability of the substitution at this position.
A2: The calculations revealed substitution of Al by Cu at the Al7 sites lowers the system by about 0.3eV according to Equation 3a. As stated in the manuscript, the frame of the Cu-substituted structure, θ-(Al, Cu)13Fe4 remains almost the same as that of the parent. This energy gain (~0.3eV/cell) originates from substituting a small fraction of Al atoms by Cu, as shown in the text. This means that the Cu substituted phase is more stable and thus can be formed easier than the parental binary compound.
Q3: The authors refer to the experimental work, where the same Al7 and Al9 positions were suggested for Cu substitution. First, in this paper, the composition of the experimental samples were selected a priori on the line toward Cu, which should model the Cu/Al substitution, but this was also not proven. Second, the results of the crystal structure refinement are given without the information, which would allow to judge the substitution as relevant; no other experimental data which could be considered as pro substitution are presented. Consequently, this publication is not suitable for verification of the calculational results in the reviewed manuscript.
A3: The experimental work in Reference 25 established the Cu substitution for Al at the Al7 and Al9 sites using a combination of X-ray diffractometry, electron microscopy and thermal analysis methods for a series of AlCu samples of different Cu contents. These experiments evidenced the high stability of the Cu-substituted crystals. Moreover, this experimental paper provided lattice parameters and coordinates of atoms for the Cu-doped crystals [25]. The experimental tools have been widely used and well-established in Physical Chemistry to determine the crystal structures. The notable difference in the number of electrons between Cu and Al allows the diffraction approaches to determine the occupation of Cu at the Al sites well, according to our Physical Chemistry. Thus, we consider that the experimental works in references 25 and 26 are reliable. The experimental results showed Cu occupations at Al7 and Al9 sites [25], which agrees well with our first-principles modeling, considering the temperature effect. Meanwhile, experiment [26] also suggested volatility of this compound under different preparation conditions.
Q4: Furthermore, there are several fragments in the text, which indicate that the authors do have difficulties with the basics of physical chemistry. Otherwise, they would not write the following statements, which are quite senseless, if not simply wrong: i) 'Due to its small solubility in Al, Fe exists in the form of Fe-containing intermetallic compounds (Fe-IMCs), including the primary θ-Al13Fe4.'; ii) 'this structure is chemically flexible and Fe/Al can be partially replaced to form ternary compounds'; iii) ' Genba and co-workers prepared Al-Fe alloy under Ar protection in a quartz (SiO2) vessel using pure Al and Fe raw materials [26]. They first heated the sample at 850 ºC and then cooled it to 750 ºC which they quenched it into water. They observed many needle-like crystals appeared on the surface of the SiO2 glass tube. Using the X-ray diffraction patterns analysis and electron probe microanalysis, they found Cu at the Fe1, Fe2, Fe3 and Al9, Al12, Al14 sites in θ-Al13Fe4 [26]'; iv) 'We present the results of the stable θ-Al76CuVII2Fe24 phase'; etc.
A4: The present work relates to multiple fields, including quantum mechanics, electronic density-functional theory, crystallography, metallurgy. It can be fairly addressed to the Physical Chemistry, where we have published several previous works (see references [3-5, 28, 35-37, 42, 49]).
Q_i) 'Due to its small solubility in Al, Fe exists in the form of Fe-containing intermetallic compounds (Fe-IMCs), including the primary θ-Al13Fe4.'
A_i: The conclusion (cited) ‘Low solubility of Fe in Al and Fe exists in the form of Fe-containing intermetallic compounds’ (end of citation) has been well established in the Metallurgy Society. We invite the reviewer to carefully read the references [1-3, 13-17, 25-28, 31-37] and others. In the last decades, removing the harmful Fe-containing intermetallic compounds in Al alloys has been a topic of intensive studies owing to both industrial applications and academic interest.
Q_ii) 'this structure is chemically flexible and Fe/Al can be partially replaced to form ternary compounds'
A_ii: This can be clearly understood from the previous experimental works and theoretical studies [22-24, and from the publications of our group [35-37].
Q_iii) ' Genba and co-workers prepared Al-Fe alloy under Ar protection in a quartz (SiO2) vessel using pure Al and Fe raw materials [26]. They first heated the sample at 850 ºC and then cooled it to 750 ºC which they quenched it into water. They observed many needle-like crystals appeared on the surface of the SiO2 glass tube. Using the X-ray diffraction patterns analysis and electron probe microanalysis, they found Cu at the Fe1, Fe2, Fe3 and Al9, Al12, Al14 sites in θ-Al13Fe4 [26]'
A_iii: We cited the work from Genba and co-workers in [26] and suggested possible explanations for the experimental results. This (and point ii) introduces a brief overview of previous experimental works and related questions to the readers.
Q_iv): ‘We present the results of the stable θ-Al76CuVII2Fe24 phase’
A_iv: Discussing this point is somewhat reasonable. Indeed, it is somewhat unusual to include atomic sites. However, this notation gives more specific information about the composition and structure of the compound [35-37], and it is in the scope of Physical Chemistry. For more transparency to the reader, we changed it into a traditional note in a systemic way.
Finally, we corrected the typos and improved the English and writing style, as Review 1 suggested.
Reviewer 2 Report
The manuscript is an interesting contribution to a better understanding of these aluminium-rich phases on the basis of ab initio simulation calculations using the VASP program package. It is shown that first principal structure and coordinate optimisation calculations with the density functional GGA reproduce the experimental values known from the literature better than LDA-based calculations. The introduction of a Hubbard correction term with U = 4eV is comprehensibly justified. In particular, the presentation of the preferential occupation of Al7 of the compound Fe4Al13 by Cu provides interesting information for the interpretation of the crystallochemical properties of this phase.
the following comments and notes on typing errors should be taken into account
Line 39: a = 8.078 Å
Line 88: α-Fe
Line 93: experimental values of the magnetic moment of Fe should be given beside the reference on calculated values.
Line 171: no quaternary phases are investigated
Line 175: use [1 0 0] direction or a-axis
Line 183: the differences between the calculated formation energies are not significant and do not allow to identify the ground state here.
Line 186 brackets missing
Line 189: the color red is not visible in Figure 3a
Line 199: Θ‘
Lines221-226: these finding are very interesting and should be discussed in more details. In respect to the atomic radii it is not straight forward to motivate longer Cu-Fe distances (probably not bondlenght) then Al-Fe distances. The same is true for shorter Cu-Fe distances compared to Fe-Fe distances.
Line 233: Please justify more clearly why the number of Fe neighbours does not matter.
Line 253: Θ-Al76Cu2Fe24
Furthermore, this used notation is not suitable. It is completely unusual to include the atomic position in the phase name.
Line 254: GGA
Line 258: s.a. line 175
Line 270: "compounds" should be replaced by "compositions"
Line 273: Figure 5a coordinate system should be given
Line 292: Figure 7: only the interesting part around the Fermi level should be presented +/- 1ev or smaller.
Line 312: The composition relation should be better explained and motivated.
Author Response
Responses to Reviewer 2
Q1: The manuscript is an interesting contribution to a better understanding of these aluminium-rich phases on the basis of ab initio simulation calculations using the VASP program package. It is shown that first principal structure and coordinate optimisation calculations with the density functional GGA reproduce the experimental values known from the literature better than LDA-based calculations. The introduction of a Hubbard correction term with U = 4eV is comprehensibly justified. In particular, the presentation of the preferential occupation of Al7 of the compound Fe4Al13 by Cu provides interesting information for the interpretation of the crystallochemical properties of this phase.
A1: We thank very much Reviewer 2 for his/her careful reading and good understanding of the present work.
Q2: the following comments and notes on typing errors should be taken into account
Line 39: a = 8.078 Å
Line 88: α-Fe
A2: We corrected the typo and missing characters.
Q3: Line 93: experimental values of the magnetic moment of Fe should be given beside the reference on calculated values.
A3: We added the experimental value of magnetic moment of Fe from the literature.
Q4: Line 171: no quaternary phases are investigated
Line 175: use [1 0 0] direction or a-axis
A4: We made the requested corrections.
Q5: Line 183: the differences between the calculated formation energies are not significant and do not allow to identify the ground state here.
A5: We improved the related text about the suggested ground state of this phase.
Q6: Line 186 brackets missing
Line 189: the color red is not visible in Figure 3a
Line 199: Θ‘
A6: We corrected the typos and changed the color in Figure 3a.
Q7: Lines221-226: these finding are very interesting and should be discussed in more details. In respect to the atomic radii it is not straight forward to motivate longer Cu-Fe distances (probably not bondlenght) then Al-Fe distances. The same is true for shorter Cu-Fe distances compared to Fe-Fe distances.
Line 233: Please justify more clearly why the number of Fe neighbours does not matter.
A7: We included more discussions about the factors, including the roles of atomic orbitals in the chemical bonding accordingly.
Q8: Line 253: Θ-Al76Cu2Fe24
Furthermore, this used notation is not suitable. It is completely unusual to include the atomic position in the phase name.
A8: We made the corresponding changes in a systematic way.
Q9: Line 254: GGA
Line 258: s.a. line 175
Line 270: "compounds" should be replaced by "compositions"
Line 273: Figure 5a coordinate system should be given
Line 292: Figure 7: only the interesting part around the Fermi level should be presented +/- 1ev or smaller.
A9: We corrected the typos, improved Figure 5a and Figure 7 accordingly.
Q10: Line 312: The composition relation should be better explained and motivated.
A10: We improved the interpretation about the composition relation and the related thermodynamical assessments in the literature.
