Near edge x-ray absorption fine structure spectroscopy study of atomic nitrogen implanted in Al2O3 by low energy N2+ bombardment

Near edge x-ray absorption fine structure spectroscopy study of atomic nitrogen implanted in Al2O3by low energy N2 ¿bombardment A. Barranco, J. P. Holgado, F. Yubero, J. P. Espino ´s, A. Martı ´n, and A. R. Gonza ´lez-Elipea) Instituto de Ciencia de Materiales de Sevilla (CSIC-Universidad de Sevilla) and Departamento de Quı ´mica Inorga ´nica, c/Ame ´rico Vespucio s/n E- 41092 Sevilla, Spain 共Received 27 October 2000; accepted 22 January 2001兲 N2 ⫹bombardment of Al2O3has been investigated by near edge x-ray absorption fine structure spectroscopy. Two kinds of species were detected and were attributed to implanted nitrogen atoms and nitride species. These results are discussed in relation to previous attributions in the literature of these species to AlNO and AlN. © 2001 American Vacuum Society. 关DOI: 10.1116/1.1355762兴 I. INTRODUCTION Low energy ion implantation in metal oxides has been widely studied because of its fundamental and applied interest for the preparation of thin films and other technological materials.1Nitrogen 共i.e., N2 ⫹兲bombardment of metal oxides has been reported to produce changes in stoichiometry by substitution of oxygen by nitrogen ions.2,3 Typically, the chemical state of this implanted nitrogen has been recognized as nitride species. Choi et al.4,5 have recently reported that more than one type of nitrogen species is formed when Al2O3is subjected to low energy (0.5⬍E⬍2 keV) ion bombardment. Based on the binding energy 共BE兲values of the x-ray photoemission spectroscopy 共XPS兲peaks of these species, Choi et al. assigned them to AlON (N 1sBE 402.5 eV兲 and Al–N (N1sBE 396 eV兲. The first type of species is formed by low energy ion bombardment (E⬇0.5keV), while both of them are formed at E⭓1 keV. Besides, a badly defined peak at 397.6 eV, which appears at intermediate values of bombarding energy, is attributed to Al–N. Very recently we have also reported these N2 ⫹bombardment processes in Al2O3.6Basically this is a continuation of previous work by Bertoti et al.7on the same subject. In both cases we have found by XPS a behavior similar to that reported by Choi et al. However, our interpretation of the higher binding energy peak substantially differs from that of those authors. After a careful analysis of XPS and ion scattering spectroscopy 共ISS兲spectra in a similar ion implantation experiment, in which peak positions, intensities and thermal stability of these species were considered, we proposed that the peak should be attributed to atomic nitrogen occupying lattice and/or vacancy positions of an altered Al2O3network.6In the same paper the appearance of a small O1speak at 537 eV upon O2 ⫹bombardment was attributed to the formation of similarly trapped species of atomic oxygen. With the aim of adding more support to this assignment we have carried out experiments by x-ray absorption spectroscopy 共XAS兲by recording the near edge fine structure 共NEXAFS兲spectrum of nitrogen implanted Al2O3samples. NEXAFS spectra of elemental atomic species of nitrogen implanted in solids are typically characterized by sharp resonance peaks.8–10 So, in principle, their identification should be easy since the spectra are characterized by narrow sharp peaks, in contrast with the broader structure of typical nitride species. II. EXPERIMENT The XPS spectra of nitrogen implanted in Al2O3were recorded on a Leybold–Heraeus LHS-10 spectrometer working in the pass energy mode at a value of 50 eV. The MgK ␣ line (hv⫽1253.6eV) was used as the excitation source. A reference for the BE scale of the spectra was taken at 284.6 eV for the C1speak of the carbon contaminating the sample. N2 ⫹bombardment in the energy range of 0.5⬍E ⬍2 keV was carried with a Leybold–Heraeus hollow cathode ion gun 共IQP-10兲. Bombardment was carried out for 30 min 共total dose of 0.9 and 5.8⫻1016 ionscm⫺2for 0.5 and 2 keV, respectively兲up to saturation of the N1ssignals. A thorough XPS investigation of this N2 ⫹bombardment process can be found in a previous paper.6After preparation the samples were exposed to air and stored without any special precaution. Apart from an increase in the intensity of the C1ssignal, no change in the shape of N1ssignals or in the relative intensity of spectrum components was detected by XPS after several weeks of exposure to air. However some attenuation of the Al2p,O1s, and N1ssignals was observed, as expected from this surface contamination by carbon. These samples were investigated by NEXAFS together with other untreated Al2O3samples that were bombarded in situ in the ultrahigh vacuum 共UHV兲chamber used for the NEXAFS experiment. No noticeable differences in the shape of the spectra were detected for the ex situ or in situ samples subjected to nitrogen bombardment at the same energies, although the intensity of the signals was slightly smaller in the former case probably due to contamination of the surface. NEXAFS spectra at the N Kedge were recorded at the PM-1 beam line at the BESSY II synchrotron in Berlin 共Germany兲. The photon energy was selected with a SX-700 plane grating monochromator and the spectra were recorded by a兲Electronic mail: [email protected] 1024 1024J. Vac. Sci. Technol. A 19„3…, MayÕJun 2001 0734-2101Õ2001Õ19„3…Õ1024Õ3Õ$18.00 ©2001 American Vacuum Society total electron yield 共TEY兲by measurement of the drain current through the sample or by channeltron detection. Except for the background line, similar shapes were recorded for the spectra by these two methods of detection. Determination of photon intensity at each energy 共i.e., I0measurement兲was done by measuring the current with a gold grid placed in front of the samples. Spectra are represented after dividing the sample signal by this reference signal. For NEXAFS analysis, the Al2O3samples consisted of thin films 共thickness ⬃0.5 ␮ m兲of this material deposited onto a metal substrate by plasma enhanced chemical vapor deposition 共CVD兲. By XPS analysis either these thin film samples or ␣ -Al2O3single crystals were used. In both cases the XPS results obtained were quite similar. Thin film samples were preferred for XAS studies because they presented enough conductivity so to avoid undesired charging effects and to enable drain current recording of the spectra. III. RESULTS AND DISCUSSION Figure 1 shows N1sphotoelectron spectra for nitrogen implanted at 0.5 and 2 keV on Al2O3thin films. These spectra are equivalent to those previously reported by Choi et al.4,5 and by ourselves6for a similar bombardment experiment on ␣ -Al2O3single crystals. By low energy bombardment a N1speak at 403.4 eV is the majority species detected. Besides, a badly defined structure appears at lower BE. Meanwhile, by bombardment at the high energy, two peaks at 403.4 and 396.4 eV are clearly detected. In this second case the asymmetry on the high BE side of the 396.4 eV peak still suggests the presence of minority species with an intermediate value of BE. These two samples were then examined by NEXAFS spectroscopy and their corresponding spectra are shown in Fig. 2. In the sample bombarded with 0.5 keV ions a sharp peak at 398 eV with a full width at half maximum 共FWHM兲of ⬃1 eV is the most noticeable feature observable in the spectrum. This signal must have originated from the same nitrogen species as that producing the N1s photoelectron band at 403.4 eV. Meanwhile, in the XAS spectrum of the sample bombarded with 2 keV ions this sharp peak, together with a much broader signal, is detected. This second signal is characterized by main maxima at around 404 and 411 eV and is similar to the N1sNEXAFS spectra of nitride species reported in many metal nitride systems11 共Note that these energy values refer to the approximate position of the maxima and not to the energies of the absorption thresholds and that, therefore, they cannot be directly compared with BEs of photoemission peaks兲. Some hints of a similar species can be also seen in the 0.5 keV spectrum. A small concentration of a species of that kind would agree with the appearance of a badly defined structure between 400 and 396 eV in the photoelectron spectrum of the 0.5 keV sample 共cf. Fig. 1兲. However, additional investigations are still necessary to properly account for these small features. Sharp resonance peaks of 1 eV width at the N Kedge absorption spectra are not compatible with N–O species. Spectra of absorbed NO on metal surfaces are characterized by ␲ *and ␴ *resonances at ⬃400 and 412 eV.12 In nonoriented samples an average weighted intensity should be expected for these two resonances, a situation that is not found in our experiment with 0.5 keV. N Kedge spectra of molecular nitrogen adsorbed on metals have also been reported.13,14 Two features at 401, sharp and intense, and at 418 eV, broad and lesser in height, characterize this type of elemental nitrogen. The absence of any hint of this second feature at 418 eV in the spectrum of the 0.5 keV ion bombardment sample eliminates any suggestion that molecular nitrogen can be the species that cause the sharp feature at 398 eV. Steric hindrance of the incorporation of such molecular species within the aluminum oxide lattice is another reason to disregard it as responsible for the signal recorded. Thus, we attribute the sharp resonance line to atomic nitrogen species located in empty positions of the lattice. A similar sharp resonance has been reported by Soriano et al. for FIG. 1. XPS spectra of Al2O3surfaces modified by N2 ⫹ion bombardment. FIG. 2. NEXAFS spectra corresponding to the two samples in Fig. 1. 1025 Barranco et al. : NEXAFS study of atomic nitrogen implanted in Al2O31025 JVST A - Vacuum, Surfaces, and Films N2 ⫹bombarded Zr and Hf or for TiN heated in O2at high temperatures.8,9 A few years ago, Jime ´nez et al. attributed the appearance of a sharp peak at 398 eV to nitrogen atoms in interstitial positions of sp3BN,10 where only by outgassing at T⬎1000K did removal of these species occur. Thus, the present NEXAFS results support our previous assignment of the two XPS peaks at 403.4 and 396.4 eV to atomic 共i.e., N兲and nitride species of nitrogen implanted in Al2O3. It seems that at low energy 共i.e., E⬃0.5keV), nitrogen ions are not able to preferentially remove oxide ions from the Al2O3lattice. However some nitrogen atoms may become incorporated within the very rigid Al2O3network, likely at interstitial or vacancy positions. At a higher energy 共i.e., E⬃2 keV) preferential removal of oxygen must occur, and their lattice positions become occupied by nitrogen in the form of nitride species 共i.e., Al–N bonds兲. This latter effect occurs without precluding the occupancy of some interstitial sites by atomic nitrogen as indicated by the appearance of the two peaks in the photoelectron spectrum. It is also interesting that due to the high stability of these implanted species of nitrogen further ex situ thermal and environmental characterization would be possible with regard to other properties such as optical, magnetic, etc. Progress is being made in our laboratory in that direction. ACKNOWLEDGMENTS The authors thank the CICYT 共Project No. MAT97-0689兲 for financial support and the HCM program of the European Union for providing access to the synchrotron light source. 1J. C. Sa ´nchez-Lo ´pez, L. Contreras, A. Ferna ´ndez, A. R. Gonza ´lez-Elipe, J. M. Martı ´n, and B. Vacher, Thin Solid Films 317, 100 共1998兲. 2J. Taylor and J. W. Rabalais, J. Chem. Phys. 75, 1735 共1981兲. 3I. Berto ´ti, R. Kelly, M. Mohai, and A. To ´th, Nucl. Instrum. Methods Phys. Res. B 80Õ81, 1219 共1993兲. 4W.-K. Choi, S.-C. Choi, H.-J. Jung, S.-K. Koh, D. Byun, and D.-W. Kum, J. Vac. Sci. Technol. A 16, 3311 共1998兲. 5W.-K. Choi, S.-K. Koh, and H.-J. Jung, J. Vac. Sci. Technol. A 17, 3362 共1999兲. 6J. P. Espino ´s, A. R. Gonza ´lez-Elipe, M. Mohai, and I. Bertoti, Surf. Interface. Anal. 共in press兲. 7I. Bertoti, R. Kelly, M. Mohai, and A. To ´th, Surf. Interface Anal. 19,291 共1992兲. 8L. Soriano, M. Abbate, J. C. Fuggle, C. Jime ´nez, J. M. Sanz, L. Gala ´n, C. Mythen, and H. A. Padmore, Surf. Sci. 281, 120 共1993兲. 9L. Soriano, M. Abbate, J. C. Fuggle, P. Prieto, C. Jime ´nez, J. M. Sanz, L. Gala ´n, and S. Hofmann, J. Vac. Sci. Technol. A 11,47共1993兲. 10I. Jime ´nez, A. F. Jankowski, L. J. Terminello, D. G. J. Sutherland, J. A. Carlisle, G. L. Doll, W. M. Tong, D. K. Shuh, and F. J. Himpsel, Phys. Rev. B 55, 12025 共1997兲. 11J. G. Chen, Surf. Sci. Rep. 30,1共1997兲. 12J. Sto ¨hr, K. Baberschke, R. Jaeger, T. Treichler, and S. Brennan, Phys. Rev. Lett. 47, 381 共1981兲. 13L. Wenzel, D. Arvanitis, R. Schlo ¨gl, M. Muhler, D. Norman, K. Baberschke, and G. Ertl, Phys. Rev. B 40, 6409 共1989兲. 14L. Wenzel, D. Arvanitis, R. Schlo ¨gl, M. Muhler, D. Norman, K. Baberschke, and G. Ertl, Phys. Scr. 41,1028共1990兲. 1026 Barranco et al. : NEXAFS study of atomic nitrogen implanted in Al2O31026 J. Vac. Sci. Technol. A, Vol. 19, No. 3, MayÕJun 2001