APPLIED PHYSICS LETTERS
VOLUME 76, NUMBER 14
3 APRIL 2000
Characteristics of the surface layer of barium strontium titanate thin films deposited by laser ablation V. Craciuna) and R. K. Singh Department of Materials Science and Engineering, University of Florida, Gainesville, Florida 32611-6400
共Received 1 November 1999; accepted for publication 6 February 2000兲 Ba0.5Sr0.5TiO3 共BST兲 thin films grown on Si by an in situ ultraviolet-assisted pulsed laser deposition 共UVPLD兲 technique exhibited significantly higher dielectric constant and refractive index values and lower leakage current densities than films grown by conventional PLD under similar conditions. X-ray photoelectron spectroscopy 共XPS兲 investigations have shown that the surface layer of the grown films contained, besides the usual BST perovskite phase, an additional phase with Ba atoms in a different chemical state. PLD grown films always exhibited larger amounts of this phase, which was homogeneously mixed with the BST phase up to several nm depth, while UVPLD grown films exhibited a much thinner 共⬃1 nm兲 and continuous layer. The relative fraction of this phase was not correlated with the amount of C atoms present on the surface. Fourier transform infrared spectroscopy did not find any BaCO3 contamination layer, which was believed to be related to this new phase. X-ray diffraction measurement showed that although PLD grown films contained less oxygen atoms, the lattice parameter was closer to the bulk value than that of UVPLD grown films. After 4 keV Ar ion sputtering for 6 min, XPS analysis revealed a small suboxide Ba peak for the PLD grown films. This finding indicates that the average Ba–O bonds are weaker in these films, likely due to the presence of oxygen vacancies. It is suggested here that this new Ba phase corresponds to a relaxed BST surface layer. © 2000 American Institute of Physics. 关S0003-6951共00兲05114-7兴
Bax Sr1⫺x TiO3 共BST兲 is one of the most studied dielectric materials for capacitors due to its very high dielectric constant.1,2 For high density integration applications or as a potential replacement of SiO2 thin films in metal-oxidesemiconductor 共MOS兲 capacitors it will be necessary to grow high quality thin films. Unfortunately, thin BST films always exhibit much lower dielectric constant values than those measured for the bulk because of the presence of fine grains,3,4 mechanical stresses,2,5 formation of interfacial layers,6 or rough interfaces.7 The formation of the interfacial layer has been of particular concern for the growth of BST films directly on a Si substrate because at the high temperatures and oxygen pressures required during the deposition, Si can easily oxidize and/or react with the grown film.6,8 Pulsed laser deposition 共PLD兲 is a technique that uses relatively low processing temperatures. It has been successfully applied for the growth of some of the highest quality BST films reported to date.4,5,7,9–11 We have shown that by using in situ ultraviolet-assisted PLD 共UVPLD兲 a significant improvement of the films’ properties with respect to those measured from conventional PLD grown films has been obtained.11 Since oxygen was introduced 5 s after the beginning of the ablation process, the interface composition and structure should be similar, regardless of the technique used. To account for the observed improvement in electrical and optical properties, a detailed investigation of the outermost surface layers of BST films grown by conventional PLD and UVPLD was performed and the results are presented here. a兲
Author to whom correspondence should be addressed; permanent address: Institute of Atomic Physics, Bucharest, Romania; electronic mail:
[email protected]
Details regarding the PLD system, UV source, and deposition conditions 共KrF laser, ⫽248 nm, fluence ⬃2 J/cm2, repetition rate 5 Hz, oxygen pressure 150 mTorr for UVPLD and 250 mTorr for PLD兲 are presented elsewhere.11 The 184 nm radiation emitted by the UV lamp 共6% of the output兲 dissociates molecular oxygen and forms ozone and atomic oxygen, which are known to enhance the oxygenation of the deposited films.12,13 It has also been shown that UV irradiation during deposition can improve the crystallinity of films grown under vacuum, without any oxygen.14,15 Moreover, the UV lamp was also used during cooling down from the deposition temperature to room temperature under 300 Torr of oxygen since UV-assisted oxygen anneals have also been shown to improve the properties of PLD grown films.16,17 The chemical composition and bonding were investigated by x-ray photoelectron spectroscopy 共XPS, Perkin Elmer 5100兲 using Mg K␣ radiation 共1253.6 eV兲 and takeoff angles of 45° and 20°. The residual pressure in the analysis chamber was lower than 5⫻10⫺9 Torr. Fourier transform infrared 共FTIR兲 spectroscopy investigations were performed with a Nicolet instrument in transmission mode. The optical properties and thickness of the films were measured by a variable angle spectroscopic ellipsometer 共VASE兲. The crystallinity was examined by x-ray diffraction 共XRD兲. The capacitance–voltage (C – V) characteristics for MOS capacitors were measured at 100 KHz using a HP4275A LCR meter with a Hg probe (diameter⬃0.03 mm2). The values of the dielectric constants estimated for different thickness BST films deposited at various substrate temperatures are displayed in Fig. 1. It is apparent that for the same deposition temperature and thickness, UVPLD grown films exhibited much higher ⑀ values. XPS spectra of
0003-6951/2000/76(14)/1932/3/$17.00 1932 © 2000 American Institute of Physics Downloaded 14 Sep 2001 to 128.113.8.180. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
Appl. Phys. Lett., Vol. 76, No. 14, 3 April 2000
FIG. 1. Dielectric constant values of BST films grown by PLD or UVPLD at different substrate temperatures.
the Ba 3d region acquired at a 45° takeoff angle of as-grown samples at different deposition temperatures are shown in Fig. 2. Each of the Ba 3d3/2 and Ba 3d5/2 peaks can be fitted by two peaks separated by around 1.5⫾0.1 eV. The lower binding-energy peaks, i.e., 794.2 and 779.0 eV and denoted Ba1, were assigned to Ba atoms in the BST perovskite phase.18–23 The peaks at 795.7 and 780.5 eV and denoted Ba2, were initially thought to be caused by Ba atoms in a decomposed barium carbonate layer.18–21 Since the formation of a carbonate layer on the surface of UV-annealed hydroxyapatite films grown by PLD has been recently reported17 this seemed quite likely to occur here as well. However, while FTIR analysis of hydroxyapatite films17 clearly showed the presence of absorption bands around 1450 cm⫺1, due to CO2⫺ 3 groups, no such bands were observed here, not even for samples grown at 450 °C, as one can see in Fig. 3. It is worth mentioning that FTIR has been recognized as one of the most sensitive techniques for surface carbonate layer identification22 and that, according to XPS analysis, the surface of this sample, as one can see in Fig. 2, mostly consisted of this new phase 共57%兲. While the relative fraction of the Ba2 peak decreased with the increase of the deposition temperature and the use of UV radiation, no such correlation was found for the C 1s peak. Due to charging during XPS investigations, the posi-
V. Craciun and R. K. Singh
1933
FIG. 3. FTIR transmittance spectrum of a BST film grown by PLD at 450 °C.
tion of the peaks was shifted from their reference values. The C 1s binding energy cannot be used as a reference in our particular case as we do not know whether this signal was caused by adventitious hydrocarbon usually present on the surface even in ultrahigh vacuum conditions, or was caused by the formation of a chemical compound. Upon assigning the Ba1 peak a binding energy corresponding to stoichiometric BST, i.e., 779.0 eV, one obtains 529.6 and 285.0 eV for the O 1s peak and C 1s peaks, respectively. The binding energy for O 1s is well within the accepted range for BST compounds,18–24 while that of C 1s is much closer to that of adsorbed hydrocarbon 共284.6 eV in our XPS system兲, than that of ⬃288.6–289.5 eV usually measured for carbon in BaCO3. 18–21 Taken at a takeoff angle of 20° in order to enhance surface contribution, the spectra for the Ba 3d region for samples grown at 550 °C are displayed in Fig. 4. One can see that while for the PLD grown samples the ratio of Ba1 to Ba2 peak intensities remained almost constant, at around ⬃41%, it increased significantly for the UVPLD grown sample, from ⬃30% to ⬃44%. This suggests that the surface layer of PLD grown film consists of a rather homogeneous
FIG. 2. Acquired and fitted XPS spectra of the Ba 3d region for 共a兲 UVPLD FIG. 4. Acquired and fitted XPS spectra of the Ba 3d region for 共a兲 UVPLD and 共b兲 PLD grown BST films at different temperatures. and 共b兲 PLD as-grown BST films at 550 °C; takeoff angle was 20°. Downloaded 14 Sep 2001 to 128.113.8.180. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
1934
Appl. Phys. Lett., Vol. 76, No. 14, 3 April 2000
V. Craciun and R. K. Singh
In summary, we have shown that the surface layer of BST films grown by laser ablation exhibits two chemically different Ba containing phases. The newly formed phase is not a barium carbonate compound which could be etched away by an acid dip but a relaxed BST phase, probably caused by oxygen vacancies and residual defects. The presence of this phase affects the electrical and optical properties of the grown films, especially for very thin films, where its relative contribution is more important. The lower content of this phase in UVPLD grown films can explain their overall better optical and electrical properties. This can also explain previous results that have shown that either annealing treatments under oxygen5,9 or even under N2, 25 or a plasma etching of the surface26 resulted in dramatic improvements of the electrical properties of BST films. Part of this research is funded by Grant No. DE-FG0595ER45533 from the U.S. Department of Energy. FIG. 5. Acquired and fitted XPS spectra of the Ba 3d region for 共a兲 UVPLD and 共b兲 PLD grown BST films at 550 °C after 6 min 4 keV Ar ion sputtering; takeoff angle was 20°.
mixture of these two phases which extends more than 4.0 nm inside the sample depth, while it is only a very thin film 共⬃1 nm兲 uniformly covering the BST phase for the UVPLD grown sample. After 6 min sputtering with 4.0 keV Ar ions, which removed accordingly to VASE measurements around 6.0 nm of material, only a very low C contamination was found on the surface, probably caused by the vacuum system itself. The XPS analysis showed that the Ti signal now evidenced the presence of a small contribution from a Ti3⫹ state caused by loss of oxygen due to ion bombardment.18,24 The Ba 3d peak of the UVPLD grown sample showed only the presence of the BST phase 共Ba1兲 as one can see in Fig. 5, confirming previous results that Ar ion bombardment at such low energies is not inducing changes in Ba oxidation state for stoichiometric BST films.18,24 However, the Ba 3d peak collected from the PLD sample showed the presence of a lower oxidation state, as one can see in Fig. 5, which suggests that for these films a fraction of the Ba–O bonds are somehow weaker, because of oxygen vacancies and residual defects. It is interesting to note that XRD investigations have shown that the lattice parameter of PLD grown films was around 3.96 nm, very close to the 3.95 nm value recorded from BST powder used to make the targets. The UVPLD grown film exhibited a significant higher value, of around 3.98 nm, which shows that these films are strained. XPS analysis indicated that PLD grown films have less oxygen atoms than UVPLD grown films, which should have resulted in films with a larger lattice parameter,4,5 quite the opposite of what was observed. A possible explanation could be that the BST film grown by PLD has relaxed near the surface region in a different phase, where Ba atoms are located in a different chemical environment than in perovskite BST.
1
J. C. Shin, J. Park, C. S. Hwang, and H. J. Kim, J. Appl. Phys. 86, 506 共1999兲. 2 T. M. Shaw, Z. Suo, M. Huang, E. Liniger, R. B. Laibowitz, and J. D. Baniecki, Appl. Phys. Lett. 75, 2129 共1999兲. 3 G. Arlt, D. Hennings, and G. de With, J. Appl. Phys. 58, 1619 共1985兲. 4 H.-F. Cheng, J. Appl. Phys. 79, 7965 共1996兲. 5 L. A. Knauss, J. M. Pond, J. S. Horwitz, D. B. Chrisey, C. H. Mueller, and R. Treece, Appl. Phys. Lett. 69, 25 共1996兲. 6 Y. Gao, A. H. Mueller, E. A. Irene, O. Auciello, A. Krauss, and J. A. Schultz, J. Vac. Sci. Technol. A 17, 1880 共1999兲. 7 L. Kinder, X. F. Zhang, I. L. Grigorov, C. Kwon, Q. X. Jia, L. Luo, and J. Zho, J. Vac. Sci. Technol. A 17, 2148 共1999兲. 8 S.-H. Nam and H.-G. Kim, Ferroelectrics 152, 79 共1994兲. 9 W. Chang, J. S. Horwitz, A. C. Carter, J. M. Pond, S. W. Kirchoefer, C. M. Gilmore, and D. B. Chrisey, Appl. Phys. Lett. 74, 1033 共1999兲. 10 S. Saha and S. B. Krupanidhi, Mater. Sci. Eng., B 57, 135 共1999兲. 11 A. Srivastava, V. Craciun, J. Howard, and R. K. Singh, Appl. Phys. Lett. 75 共1999兲. 12 G. Koren, A. Gupta, and R. J. Baseman, Appl. Phys. Lett. 54, 1920 共1989兲. 13 H. Tabata, T. Kawai, and S. Kawai, Appl. Phys. Lett. 58, 1443 共1991兲. 14 V. Craciun, J. Howard, and R. K. Singh, presented at the Materials Research Society Spring Meeting, San Francisco, 6–10 April, 1999. 15 H. Wengenmair, J. W. Gerlach, U. Preckwinkel, B. Stritzker, and B. Rauschenbach, Appl. Surf. Sci. 99, 313 共1996兲. 16 V. Craciun, D. Craciun, and I. W. Boyd, Electron. Lett. 34, 1527 共1998兲. 17 V. Craciun, I. W. Boyd, D. Craciun, P. Andreazza, and J. Perriere, J. Appl. Phys. 85, 8410 共1999兲. 18 C. Miot, E. Husson, C. Proust, R. Erre, and J. P. Coutures, J. Mater. Res. 12, 2388 共1997兲. 19 C. Miot, C. Proustz, E. Husson, R. Erre, G. Blondiaux, J. M. Beny, and J. P. Coutures, Mater. Sci. Eng., B 45, 17 共1997兲. 20 Y. Fujisaki, Y. Shimamoto, and Y. Matsui, Jpn. J. Appl. Phys., Part 2 38, L52 共1999兲. 21 M. I. Sosulnikov and Yu. A. Teterin, J. Electron Spectrosc. Relat. Phenom. 59, 101 共1992兲. 22 M. C. B. Lopez, G. Fourlaris, B. Rand, and F. L. Riley, J. Am. Ceram. Soc. 82, 1777 共1999兲. 23 S. M. Mukhopadhyay and T. C. S. Chen, J. Mater. Res. 10, 1502 共1995兲. 24 D. Leinen, A. Fernandez, J. P. Espinos, and A. R. Gonzalez-Elipe, Appl. Phys. A: Mater. Sci. Process. 63, 237 共1996兲. 25 C. S. Hwang, S. O. Park, H. Cho, C. S. Kang, H. Kang, S. I. Lee, and M. Y. Lee, Appl. Phys. Lett. 67, 2819 共1995兲. 26 H.-J. Cho, S. Oh, C. S. Kang, C. S. Hwang, B. T. Lee, K. H. Lee, H. Horii, S. I. Lee, and M. Y. Lee, Appl. Phys. Lett. 71, 3221 共1997兲.
Downloaded 14 Sep 2001 to 128.113.8.180. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
