JOM, Vol. 65, No. 11, 2013

DOI: 10.1007/s11837-013-0741-x Ó 2013 TMS

Oxidation of Cr2AlC (0001): Insights from Ab Initio Calculations NENG LI,1,2,3 RIDWAN SAKIDJA,1 and WAI-YIM CHING2 1.—Department of Physics and Astronomy, University of Missouri-Kansas City, Kansas City, MO 64110, USA. 2.—Center for Photovoltaics and Solar Energy, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, P. R. China. 3.—e-mail: lineng@ umkc.edu

We performed an investigation of the initial stage of oxidation onto a relevant Cr2AlC (0001) surface by ab initio calculations. For the most energetically stable Al-terminated Cr2AlC (0001) surface, a detailed model describing the oxygen-surface interaction is developed by exploring the adsorption energetics. Based on the evaluation of the energetics and the structural properties of the atomistic models generated, the results point to an initial stage of the Cr2AlC (0001) surface oxidation with some similarities with those observed in the Al (111) layer. Our findings on the bonding mechanism of single O adsorption atoms of the surface may lead to further alloying strategies to enhance oxidation resistance in a wide range of refractory-metal-based MAX phases.

INTRODUCTION The MAX phases, as one group of inherent nanolaminated compounds, have the general formula Mn+1AXn, where n = 1, 2, or 3; M is an early transition metal; A is an A-group element; and X is either C or N. The A-type atoms (e.g., Al or Si) reside on their own plane within the MAX crystal structure, whereas the M and C atoms are clustered on separate layers (see Fig. 1). In recent years, MAX phases attracted a great deal of attention from researchers, in both the experimental and simulation fields.1–4 In the experimental aspect, various properties of MAX phases have been investigated.2,3 In the simulation aspect, the studies are mainly concerned with the atomic and electronic structures, phase stability,5–10 and elastic and mechanical properties of bulk MAX phases.11–19 However, the atomistic mechanisms of the oxidation process for the MAX phases are not fully understood despite the extensive oxidation and corrosion studies at the macroscopic level and the fact that there is a wide range of oxidation resistance levels exhibited by the MAX phases.20–22 For this study, we focused on Cr2AlC as one example of oxidation-resistant MAX phases that is also a promising candidate for high-temperature structural applications. We performed density functional theory (DFT) calculations on a 3 9 3 supercell structure so as to simulate the adsorption process of an oxygen atom onto the MAX Al-terminated surface

(Published online September 5, 2013)

both at the ground state (T = 0 K) as well as at room temperature (T = 298 K) via ab initio molecular dynamics (AIMD) as implemented in the VASP Vienna ab initio simulation package (VASP).23–25 The main objective of this work is to gain insights into the atomistic mechanisms and in particular into the site preference on the (0001) Al layer surface of Cr2AlC as the first step toward the formation of the Al-oxide monolayer. We chose the Al-terminated Cr2AlC (0001) surface based on the previous theoretical studies26,27 with a 1 9 1 9 1 and 2 9 2 9 1 supercells showing that it has the lowest surface energy. The DFT study comprises the adsorption of single O atoms, through the sampling of several adsorption sites. Based on the evaluation of the energetics and the structural properties of the atomistic models generated, the results here reported delineate a consistent picture of the initial stage of the surface oxidation. METHODOLOGY To perform efficient surface structure and O adsorption simulations on Cr2AlC (0001), we employed a DFT-based, plane-wave, projector-augmented wave method as implemented in VASP. The structural relaxation is carried out by using VASP. For structural relaxation, the PAW-PBE potential28 with generalized gradient approximation as supplied in the VASP is used for the exchange-correlation

1487

1488

Li, Sakidja, and Ching

favorable position. The oxygen-surface interactions on the (3 9 3)/Al-terminated Cr2AlC (0001) surface are performed by exploring thoroughly the adsorption energetics (Eads), which is a key quantity in predicting the adhesive property of an adsorption system. The Eads, which is defined as the reversible energy required to separate an adsorption system into a bare surface and an adsorbed atom, can be expressed by subtracting the sum of the total energy of the half-optimized O molecule (1/2EO2 ) and the bare surface (Esurf) from the total energy of the adsorption system (Eads)30 DEads ¼ ðEads  Esurf  1=2EO2 Þ

Fig. 1. Relaxed structure of (3 9 3)/Al-terminated-Cr2AlC (0001) surface: (a) side view and (b) top view.

potential. We used a cutoff energy of 400 eV, a relatively high accuracy for the ground state electronic convergence criterion (105 eV), and force conver˚ ). The stress level of the final gence limit (102 eV/A equilibrium structure is less than 0.1 GPa. The relaxation of the present model imposes no restrictions on the volume and lattice vectors of a periodic supercell. The AIMD simulations on both a single oxygen and an oxygen molecule as implemented in VASP29 were performed by means of an NVT (constant volume and temperature) ensemble with each ionic MD step of 2 fs for duration up to 3 ps. RESULTS AND DISCUSSION The O atom is initially placed on four different sites: top (T), bridge (B), fcc-hollow (H1), and hcphollow (H2) sites of the relaxed Al-terminated Cr2AlC (0001) surface. The surface model is composed of 13 layers of 3 9 3 9 1 supercell separated ˚ . The first Cr layer is positioned by a vacuum of 15 A underneath the Al layer, and the C layer is sandwiched between the two consecutive Cr layers forming layered Cr-C clusters. The pattern is then repeated back to the Al layer (see Fig. 1). Our use of ‘‘hcp-hollow’’ as well as of ‘‘fcc-hollow’’ terms here is only meant for convenience so that we can compare the results with those from the hcp-hollow and fcchollow sites on the Al(111) surface. Thus, the layer sequence of the ‘‘fcc-like’’ ABC for the MAX phase implies the ‘‘A’’ layer is the aluminum layer, ‘‘B’’ layer is the chromium layer, and ‘‘C’’ is the carbon layer (see Fig. 2). The adsorbed structures are also relaxed, and the absorbed O atom comes to its

(1)

where Eads and Esurf are the total energies of the final adsorbate system and the bare (3 9 3)/ Cr2AlC(0001) surface, respectively, and EO2 is the total energy of a free O2 molecule in its ground ˚ . With this sign state, computed in a box of side 20 A convention, a positive value of the adsorption energy corresponds to the energetic gain of an exothermic reaction. In general, a negative Eads indicates that the oxygen adsorption is exothermic, and thus, the adsorption system is energetically stable. The relative comparison quantitatively between calculated Eads for different atomic geometries of the same adsorption system is the focus of our study. The adsorption energy of the corresponding geometrical structures is presented in Fig. 3 and listed in Table I. One can clearly see that from Fig. 3, the fcc-hollow (H1) site, is the most stable binding site for O with an adsorption energy of 5.696 eV and ˚ for Al-O. The an average bond length of 2.09 A second most favorable adsorption energy, 5.233 eV of O is for the hcp-hollow (H2) site. The top (T site) and bridge sites (B site) exhibit much less favorable adsorption energy of 3.179 eV and 3.292 eV, respectively. These sites may be considered transition states for the O to diffuse on the Cr2AlC (0001) surface between the two more preferred H1 and H2 sites. Similar to that observed in the MAX phase, these two sites are the more favorable sites on the Al layer despite the fact that the atom types constructing the ‘‘B’’ and ‘‘C’’ layer structure are not made of Al. In the case of fcc Al, the fcc-hollow site is actually the one with the lowest binding energy followed by the hcp-site with the slightly lower biding energy. In addition, the magnitude of the binding energy is 1 eV  2 eV higher (e.g., fcc site = 7.65 eV31) than our results for the MAX phase. To some extent this can be explained by the nature of the bonding mechanism that is within the MAX phase. The interlayer directional bonding between the A layer and the (M + X) layer is typically weaker than the strong MX bonds that exist within the M + X clustered layer.1 As a result, the A layer’s response to the oxidation, at least at the initial stage, may be quite similar to that observed on Al(111), meaning that the oxidation response is

Oxidation of Cr2AlC (0001): Insights from Ab Initio Calculations

1489

Fig. 2. Illustration of adsorption of an oxygen atom on (3 9 3)/Al-terminated-Cr2AlC (0001) surface: (a) top (T) site, (b) bridge (B) site, (c) fcchollow (H1) site, and (d) hcp-hollow (H2) site.

Table I. Adsorption energy of an oxygen atom onto different sites at (3 3 3)/Al-terminated Cr2AlC (0001) surface Adsorption site Eads (eV)

Fig. 3. Adsorption energy of one O atom adsorption on (3 9 3)/Alterminated Cr2AlC (0001) surface for different sites: top (T) site, bridge (B) site, fcc-hollow (H1) site, and hcp-hollow (H2) site.

mostly governed by the top Al layer. Nevertheless, there is indeed a difference in terms of the nature of the interlayer bonding between the fcc and the layered MAX phases as there is still directional bonding between the two types of layers (i.e., M-Al and M-C) that is expected to be much stronger than that of the Al-Al interlayer metallic bonding present within the fcc metal. It is reasonable to assume that as a consequence the Al-O binding energy on the MAX surface should be less in magnitude than that of the Al-O on Al-fcc. Nevertheless, this suggests there is a linkage between the oxidation responses of the two types of phases for the first time. In

Top

Bridge

hcphollow

fcchollow

3.179

3.292

5.233

5.696

addition, the possible existence of partly ionic bonding between Cr and O may not be neglected and may quite possibly skew the site preference toward the hcp-hollow (H2) site. As a further verification on the adsorption process, we carried out AIMD for both a single oxygen atom and a single O2 molecular adsorption using a (3 9 3) unit cell with 117 atoms at 700 K. The single-oxygen atom as well as O2 molecule was placed on top of the Al atom at the initial configuration on each case, and the initial velocities were randomly assigned according to a Maxwell–Boltzmann distribution at room temperature. The final configurations after the 3-ps simulation both types of shows the oxygen species favorably reside on the same hcp-hollow site (see Fig. 4) which is different with predicted by the DFT calculation at the ground state, because the final configurations depends on the temperature in AIMD simulation. Certainly, further work is required to fully assess the effect of higher temperature exposure on the dynamics site preferences and at a longer time, but the current finding suggests that some similarities between the initial stage of oxygen absorption of Al(0001) in

1490

Li, Sakidja, and Ching

surfaced in a Cr-based Cr2AlC MAX phase. The O strongly bonds to the surface Al and formed Al-O bond whose mechanisms appear to be quite similar to that observed at the initial stage of oxidation on the Al (111) surface of fcc Al. The oxidation of the Cr2AlC (0001) surface is a complex reaction involving many atomistic processes, and thus, this study is an initial step to elucidate some aspects of the oxidation reaction for the refractory-metal-based MAX in general. ACKNOWLEDGEMENTS This work is supported by U.S. DOE-NETL, the award FE0005865. This research used the resources of the National Energy Research Scientific Computing Center supported by DOE under Contract No. DE-AC03-76SF00098. REFERENCES

Fig. 4. Plan view of the AIMD (NVT) simulation results on (a) single O atom and (b) and (c) O2 molecule adsorption on Al-terminated Cr2AlC (0001) surface: Temperature is at (a) 700 K, (b) 700 K, and (c) 1273 K, with the duration is 3 ps with an ionic step of 2 fs.

MAX phase and Al(111) fcc can indeed be drawn. This can potentially improve the alloying strategy needed to address the lack of oxidation resistance in other Al-containing MAX phases (e.g., Nb-based MAX). SUMMARY AND CONCLUSIONS In this article, we have evaluated the adsorption process of a single O atom onto the Al layered

1. M.W. Barsoum and M. Radovic, Annu. Rev. Mater. Res. 41, 195–227 (2011). 2. Z. Sun, Int. Mater. Rev. 56, 143–166 (2011). 3. J. Wang and Y. Zhou, Annu. Rev. Mater. Res. 39, 415–443 (2009). 4. P. Eklund, M. Beckers, U. Jansson, H. Ho¨gberg, and L. Hultman, Thin Solid Films 518, 1851–1878 (2010). 5. C. Li, Z. Wang, and C. Wang, Intermetallics 33, 105–112 (2013). 6. J. Frodelius, J. Lu, J. Jensen, D. Paul, L. Hultman, and P. Eklund, J. Eur. Ceram. Soc. 33, 375–382 (2013). 7. M. Dahlqvist, B. Alling, I.A. Abrikosov, and J. Rose´n, Phys. Rev. B 81, 024111 (2010). 8. J. Wang, Y. Zhou, T. Liao, J. Zhang, and Z. Lin, Scripta Mater. 58, 227–230 (2008). 9. J. Wang, J. Wang, Y. Zhou, and C. Hu, Acta Mater. 56, 1511–1518 (2008). 10. Y. Mo, P. Rulis, and W.Y. Ching, Phys. Rev. B 86, 165122 (2012). 11. S. Cui, D. Wei, H. Hu, W. Feng, and Z. Gong, J. Solid State Chem. 191, 147–152 (2012). 12. M.G. Brik, N.M. Avram, and C.N. Avram, Comp. Mater. Sci. 63, 227–231 (2012). 13. M.A. Ghebouli, B. Ghebouli, M. Fatmi, and A. Bouhemadou, Intermetallics 19, 1936–1942 (2011). 14. S. Cui, W. Feng, H. Hu, Z. Lv, G. Zhang, and Z. Gong, Solid State Commun. 151, 491–494 (2011). 15. Y. Bai, X. He, M. Li, Y. Sun, C. Zhu, and Y. Li, Solid State Sci. 12, 144–147 (2010). 16. M.B. Kanoun, S. Goumri-Said, and A.H. Reshak, Comput. Mater. Sci. 47, 491–500 (2009). 17. A. Bouhemadou, Solid State Sci. 11, 1875–1881 (2009). 18. Z. Sun, S. Li, R. Ahuja, and J.M. Schneider, Solid State Comm. 129, 589–592 (2004). 19. W.-Y. Ching, Y. Mo, S. Aryal, and P. Rulis, J. Am. Ceram. Soc. 96, 2292–2297 (2013). 20. S. Li, X. Chen, Y. Zhou, and G. Song, Ceram. Int. 39, 2715– 2721 (2013). 21. Z.J. Lin, M.S. Li, J.Y. Wang, and Y.C. Zhou, Acta Mater. 55, 6182–6191 (2007). 22. D.B. Lee, T.D. Nguyen, J.H. Han, and S.W. Park, Corros. Sci. 49, 3926–3934 (2007). 23. G. Kresse and J. Furthmu¨ller, Comput. Mater. Sci. 6, 15–50 (1996). 24. G. Kresse and J. Furthmu¨ller, Phys. Rev. B 54, 11169–11186 (1996). 25. G. Kresse and J. Hafner, Phys. Rev. B 47, 558–561 (1993). 26. W. Jiemin, W. Jingyang, and Z. Yanchun, J. Phys. Condens. Mat. 20, 225006 (2008). 27. Z. Sun and R. Ahuja, Appl. Phys. Lett. 88, 161913 (2006).

Oxidation of Cr2AlC (0001): Insights from Ab Initio Calculations 28. J.P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865–3868 (1996). 29. J. Hafner, J. Comput. Chem. 29, 2044–2078 (2008).

1491

30. C.R. Ashman and G. Pennington, Phys. Rev. B 80, 085318 (2009). 31. C. Lanthony, J.M. Duce´re´, M.D. Rouhani, A. Hemeryck, A. Este`ve, and C. Rossi, J. Chem. Phys. 137, 094707 (2012).