Oxygen incorporation in Ti2AlC thin films

7 downloads 0 Views 210KB Size Report
Rosen, L. Ryves, P. O. Å. Persson, and M. M. M. Bilek, J. Appl. Phys. 101, 056101 2007. 9G. Kresse and J. Furthmüller, Comput. Mater. Sci. 6, 15 1996. 10J.
APPLIED PHYSICS LETTERS 92, 064102 共2008兲

Oxygen incorporation in Ti2AlC thin films J. Rosen,1,a兲 P. O. Å. Persson,1 M. Ionescu,2 A. Kondyurin,1 D. R. McKenzie,1 and M. M. M. Bilek1 1

School of Physics, the University of Sydney, New South Wales 2006, Australia ANSTO, Lucas Heights, New South Wales 2234, Australia

2

共Received 2 November 2007; accepted 8 January 2008; published online 11 February 2008兲 Thin films of Ti2AlC MAX phase have been deposited using a multiple cathode pulsed cathodic arc. Evidence for substantial oxygen incorporation in the MAX phase is presented, likely originating from residual gas present in the vacuum chamber during deposition. The characteristic MAX phase crystal structure is maintained, in agreement with ab initio calculations, supporting substitutional O in C lattice positions. On the basis of these results, we propose the existence of a MAX phase-like material with material properties tuned by the incorporation of oxygen. Additionally, possible unintentional O incorporation in previously reported MAX phase materials is suggested. © 2008 American Institute of Physics. 关DOI: 10.1063/1.2838456兴 The M n+1AXn phases are a family of nanolaminate ternary carbides or nitrides 共X兲, where M is an early transition metal and A is an element usually belonging to group IIIA or IVA.1,2 These compounds are attracting increasing attention due to their unique combination of metallic and ceramic properties.3 MAX phases have previously been synthesized primarily by bulk synthesis techniques and as thin films using chemical vapor deposition. Over the last few years, physical vapor deposition methods, primarily magnetron sputtering, have also been used 共see, e.g., Ref. 4兲, offering the advantage of comparatively low temperature synthesis. The MAX phase properties stem from the highly anisotropic hexagonal crystal structure, wherein M and X atoms form octahedral edge sharing building blocks interleaved by layers of A atoms. Hence, it is expected that structural defects could be crucial for retaining these properties. Also, by changing the local chemical environment for atoms in the structure, impurities would be expected to alter material properties and in some cases trigger phase changes. Vacuum based deposition techniques are often characterized by the presence of residual gas, with the major constituent being H2O. The resulting H incorporation in oxides has been shown to be crucial for film composition and structure, and hence for the film properties.5 The few reports on defects in MAX phase thin films mainly concern structural defects, see, e.g., Ref. 6. To date, there are no studies focusing on impurity incorporation during material synthesis. Nevertheless, Emmerlich et al. investigated the thermal stability of Ti3SiC2 thin films,7 and observed O indiffusion 共from residual gas in the evacuated furnace兲 in a defective MAX phase structure, resulting in O incorporation of around 10 at. %. Based on previous studies,5,7 oxygen and/or hydrogen incorporation may be expected to occur also in as deposited MAX phase thin films. We have, therefore, performed a detailed study on impurity incorporation in arc deposited Ti2AlC. This method was recently demonstrated to be a method for MAX phase synthesis.8 We show evidence for substantial oxygen incorporation without disruption to the characteristic MAX phase structure. The findings are of importance firstly because they identify a MAX-phase-like maa兲

Electronic mail: [email protected].

terial which contains a large concentration of oxygen, and secondly, because they point toward the possibility of tailoring the unique MAX phase properties, from those of the “pure” carbides/nitrides to those of the corresponding oxides. Furthermore, our results are suggestive of possible unintentional O incorporation in previously reported MAX phase materials. The thin films were deposited using a high current pulsed cathodic arc, operating at a base pressure of 1 ⫻ 10−6 mbar. Three center triggered cathodes of Ti, C, and Al were used in alternating mode, delivering plasma pulses of 350, 850, and 300 ␮s, respectively, in the pulse ratio 10:1:10. For more details on experimental parameters and optimization thereof, see Ref. 8. Prior to deposition, sapphire 共0001兲 substrates were rinsed in ethanol and then degassed in the vacuum chamber at 900 ° C for ⬃5 min. Under isothermal conditions, a layer of ⬃10 nm TiC was then deposited onto the substrate, with subsequent MAX phase deposition at a rate of ⬃5 Å s−1. The structure of the film was characterized with x-ray diffraction 共XRD兲 and transmission electron microscopy 共TEM兲. For XRD analysis, both ␪-2␪ and 1° grazing incidence 共GI兲 scans were performed using Cu K␣ radiation in a Siemens D5000 diffractometer. For TEM analysis, crosssectional specimens were prepared using mechanical grinding and polishing followed by argon ion milling. Images were recorded using a JEOL 3000F instrument, operated at 300 kV. The chemical composition of the film was obtained by Rutherford backscattering 共RBS兲 in combination with elastic recoil detection analysis 共ERDA兲. He1+ ions of 1.8 MeV were used for the measurement, with the ion flux incident on the sample surface at an angle of 68° relative to the normal sample. Furthermore, Fourier transform infrared spectroscopy 共FTIR兲 was performed for structure/ composition analysis including bonding information. The angle of the incident beam was 70°, and 1000 scans were taken at a resolution of 4 cm−1. In order to study the effect of the impurity incorporation at a range of sites with atomic resolution, ab initio calculations based on density functional theory were performed using a plane wave-pseudopotential formalism as implemented in the VASP software.9 The PW91 generalized gradient approximation was used together with a plane wave cutoff of 500 eV and a Gamma centered grid of

0003-6951/2008/92共6兲/064102/3/$23.00 92, 064102-1 © 2008 American Institute of Physics Downloaded 14 Mar 2008 to 130.236.169.68. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

064102-2

Rosen et al.

FIG. 1. XRD ␪-2␪ scan revealing epitaxial film growth according to the presence of the Ti2AlC 共000l兲 peaks.

7 ⫻ 7 ⫻ 5 k points. The calculations were performed for a super cell of 2 ⫻ 2 unit cells of the Ti2AlC MAX phase. XRD analysis of the thin film is shown in Fig. 1. Epitaxial growth of the Ti2AlC MAX phase is evident from the ␪-2␪ diffractogram. A GIXRD scan and a bare substrate scan 共not shown here兲 excluded the presence of polycrystalline material and identified the peaks not originating from the film. The minor peaks around 36° and 76° are TiC peaks originating from the interface layer. The TEM micrograph in Fig. 2共a兲, shows clearly the aligned atomic planes in the ⬃100 nm MAX phase film, the interface layer in the middle, and the Al2O3 substrate at the bottom. By inspection of the images, large MAX phase grains can be identified. The higher magnification image in Fig. 2共b兲 shows the nanolaminated hexagonal nature of the Ti2AlC MAX phase. A profile showing the film composition as a function of depth in the film, obtained by RBS, is shown in Fig. 3. Both the Ti and the Al concentrations are comparatively homogeneous throughout the film, with an average Al/ Ti ratio of 0.49. The rightmost points in the graph show a sharp increase in the C concentration up to a Ti/ C ratio of 1.0, originating from the TiC seed layer. High levels of oxygen can be observed throughout the film, varying between ⬃5 and ⬃25 at. %. Furthermore, the O concentration shows a strong negative correlation with the C concentration such that an increase/ decrease in the former results in the opposite change of the latter. Results from ERDA show a hydrogen concentration of less than 0.5 at. %, which may be explained by the high ion energies associated with cathodic arc deposition and previously shown to reduce H incorporation.10 More detailed information on the O and C correlation is contained in the FTIR spectra shown in Fig. 4. Since the 1040 cm−1 line from the Al2O3 substrate is not observed, we infer that the line spectrum originates from the film. Strong lines are found at

Appl. Phys. Lett. 92, 064102 共2008兲

FIG. 3. Depth profile of film composition obtained with RBS.

633, 568, 438, and 421 cm−1, and a weak line at 1190 cm−1. The weak line is possibly due to C–O vibrations11 while the four strong lines here can all be attributed to vibrations in Ti–O structures.12 The line at 568 cm−1 could also include a contribution from Al–O vibrations as seen in deposited aluminum oxide compounds.13 The experimental data we have presented above show that the Ti2AlC MAX phase can incorporate substantial amounts of oxygen while retaining the characteristic hexagonal structure, and that the presence of this oxygen cannot be detected in XRD and TEM image studies. The strong correlation between O and C in the RBS composition profile suggests that the incorporated oxygen atoms occupy the carbon lattice sites. The substitution of oxygen on the carbon sites is also supported by the fact that C is bonded only to Ti in the pure MAX phase and FTIR data show strong features associated with Ti–O vibrations. This interpretation is supported further in the literature by another electron microscopy study 共of a similar material兲 including composition and bond analysis carried out using electron energy loss spectroscopy.14 The material is comparable to the so-called metallic oxycarbides, MCxOy 共M = early transition metal兲.15 For the carbide regime when M = Ti, the films crystallize in a TiC B1 NaCl-type crystal structure. Since TiC is isomorphic with TiO, and due to the high chemical affinity of Ti to O, O as an impurity may substitute C over a wide range of concentrations.15,16 To test this idea further, we conducted ab initio calculations. These showed that the hexagonal crystal structure is fully retained for configurations with 1, 2, or 3 O atoms in C positions within the chosen super cell. The corresponding O concentrations were 6.25, 12.5, and 18.75 at. % and the maximum changes in calculated cell pa-

FIG. 2. Cross-sectional TEM images showing the nanolaminated hexagonal structure of the Ti2AlC film. View 共a兲 includes also the TiC interlayer and the substrate, as indicated by the arrows. FIG. 4. FTIR reflection spectrum of the film. Downloaded 14 Mar 2008 to 130.236.169.68. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

064102-3

rameters were −0.2%, −0.8%, and −1.2% 共a axis兲 and −0.5%, −0.7%, and −0.5% 共c axis兲, respectively. The oxygen found in our film most likely originates from residual gas, primarily water, in contrast to Persson et al. who demonstrates Al2O3 substrate decomposition as a source for O incorporation.14 Specific sources may be the comparatively porous graphite cathode with possible inclusions of water, and/or water desorbed from the chamber walls upon heating during deposition. Because of the intermittent operation of the cathodic arc used, no steady state is reached and fluctuations in O incorporation, as we observed, can be expected. In conclusion, we have shown significant oxygen incorporation in the Ti2AlC MAX phase, while retaining the characteristic hexagonal crystal structure. Furthermore, our results suggest that the O impurities occupy the C lattice sites. Hence, a MAX phase with X = C or O has been identified. The results also suggest that unintentional O incorporation in previously reported MAX phase materials is possible, especially in cases where vacuum methods were used and the characterization was limited to methods which are insensitive to composition, such as XRD and TEM. V. H. Nowotny, Prog. Solid State Chem. 5, 27 共1971兲.

1

Appl. Phys. Lett. 92, 064102 共2008兲

Rosen et al.

M. W. Barsoum, Prog. Solid State Chem. 28, 201 共2000兲. M. W. Barsoum and T. El-Raghy, Am. Sci. 89, 334 共2001兲. 4 O. Wilhelmsson, J.-P. Palmquist, T. Nyberg, and U. Jansson, Appl. Phys. Lett. 85, 1066 共2004兲. 5 J. M. Schneider, K. Larsson, J. Lu, E. Olsson, and B. Hjörvarsson, Appl. Phys. Lett. 80, 1144 共2002兲. 6 P. O. Å. Persson, S. Kodambaka, I. Petrov, and L. Hultman, Acta Mater. 55, 4401 共2007兲. 7 J. Emmerlich, D. Music, P. Eklund, O. Wilhelmsson, U. Jansson, J. M. Schneider, H. Högberg, and L. Hultman, Acta Mater. 55, 1479 共2007兲. 8 J. Rosen, L. Ryves, P. O. Å. Persson, and M. M. M. Bilek, J. Appl. Phys. 101, 056101 共2007兲. 9 G. Kresse and J. Furthmüller, Comput. Mater. Sci. 6, 15 共1996兲. 10 J. Rosen, E. Widenkvist, K. Larsson, U. Kreissig, S. Mráz, C. Martinez, D. Music, and J. M. Schneider, Appl. Phys. Lett. 88, 191905 共2006兲. 11 G. Socrates, Infrared Characteristic Group Frequencies 共Wiley, Chichester, 1994兲. 12 A. Damin, F. X. L. Xamena, C. Lamberti, B. Civalleri, C. M. ZicovichWilson, and A. Zecchina, J. Phys. Chem. B 108, 1328 共2004兲. 13 R. A. Nyquist, The Handbook of Infrared and Raman Spectra of Inorganic Compunds and Organic Salts 共Academic, San Diego, 1997兲. 14 P. O. Å. Persson, J. Rosen, C. Höglund, D. R. McKenzie, and M. M. M. Bilek, J. Appl. Phys. 共submitted兲. 15 A. C. Fernandes, F. Vaz, L. Rebouta, A. Pinto, E. Alves, N. M. G. Parreira, P. Goudeau, E. Lê Bourhis, and J. P. Rivière, Surf. Coat. Technol. 201, 5587 共2007兲. 16 H. O. Pierson, Handbook of Refractory Carbides and Nitrides 共Noyes, Westwood, 1996兲. 2 3

Downloaded 14 Mar 2008 to 130.236.169.68. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp