DOI: 10.1021/cg901165j
Lattice Guiding for Low Temperature Crystallization of Rhombohedral Perovskite-Structured Oxide Thin Films
2010, Vol. 10 761–764
Sharath Sriram,*,† Madhu Bhaskaran,† David R. G. Mitchell,‡,§ and Arnan Mitchell† †
Microelectronics and Materials Technology Centre and Platform Technologies Research Institute, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, Victoria 3001, Australia and ‡Institute of Materials Engineering, Australian Nuclear Science and Technology Organisation (ANSTO), PMB 1, Menai, New South Wales 2234, Australia. §Present address: Electron Microscopy Unit, University of Sydney, Sydney, New South Wales 2006, Australia. Received September 22, 2009; Revised Manuscript Received December 7, 2009
ABSTRACT: Low temperature crystallization of complex oxide thin films has proved to be a challenge with deposition of such materials often carried out at elevated temperatures in excess of 600 °C. This article demonstrates one of the first instances of deposition of preferentially oriented strontium-doped lead zirconate titanate thin films at a relatively low temperature of 300 °C. This was achieved by carrying out deposition on gold-coated silicon substrates which exert a guiding influence on thin film growth due to similarity in lattice parameters. The microstructure and preferential orientations were studied using high resolution transmission electron microscopy and X-ray diffraction. These results illustrated the pronounced texture in the deposited thin films due to lattice guiding, with crystal structure simulations also verifying the guiding effect.
Introduction In the pursuit of enhanced functionality and miniaturization, high performance thin film coatings are increasingly incorporated into device designs. Oxides form a very large category of materials used for processes reliant on functionality. Oxide compounds can exhibit a wide variety of properties such as semiconduction, piezoelectricity, ferroelectricity, gas sensitivity, etc. Oxides can exhibit one or more functional properties, with properties often enhanced in thin films, leading to the widespread use of the terminology “multifunctional oxide thin films”. The ability to exhibit multifunctional behavior by an oxide is determined by its chemical composition and crystal structure. With thin film deposition often done from pure source material of desired composition, the resulting structure is the primary concern (although the stoichiometry of the thin films will need to be verified). Most high performance multifunctional oxides often have complex crystal structure symmetries, such as tetragonal or rhombohedral phases.1 This structure is often collectively termed “perovskite”.2,3 Such compounds generally have the ABO3 chemical structure. An example of such a multifunctional oxide with perovskite structure is lead titanate (PbTiO3) which exhibits piezoelectric, ferroelectric, and electro-optic properties. In this case, lead (Pb) is the A-site compound and titanium (Ti) is the B-site compound. The properties exhibited by such multifunctional oxide compounds can be further enhanced by the addition of dopants. These dopants can replace either a small percentage of A- or B-site atoms or both. This results in “complex multifunctional oxides”. One of the most popular compounds in this category, with a B-site substituent is lead zirconate titanate [PZT: Pb(Zr1-xTix)O3], in which the ferroelectric and piezoelectric properties are enhanced, resulting also in a more stable compound. It has been demonstrated that the perovskite-structured
piezoelectrics (such as lead zirconate titanate and barium strontium titanate) exhibit much higher levels of piezoelectric response than the simpler asymmetric structures (e.g., quartz and zinc oxide).4 Moreover, the increase in piezoelectric response is apparent in compounds with either A- or B-site substituents or both;barium strontium titanate (Sr as an A-site dopant) provides better Curie point control over barium titanate5 and lead zirconate titanate (Zr as an B-site dopant) has better piezoelectric properties than lead titanate.6 For all the potential demonstrated by functional oxides, incorporation of these materials into electronic devices has proved to be a challenge. The major bottleneck in realizing devices is the synthesis conditions for oxide thin films which demand high temperature deposition and/or processing. This article reports in detail on the first instance of low temperature deposition of perovskite structured thin films. PSZT thin films with a composition of (Pb0.92Sr0.08)(Zr0.65Ti0.35)O3 were deposited by RF magnetron sputtering on gold-coated silicon substrates. The PZT family of compounds are conventionally deposited at temperatures greater than 600 °C on platinum-coated silicon substrates.7,8 In order to improve the guiding effect from the bottom metal layer, gold was chosen as its lattice spacings closely match those of PSZT.9,10 Gold is also inert in an oxygen atmosphere (the sputtering gas consists of 10% oxygen in argon). Gold readily reacts with silicon at low temperatures (the gold-silicon eutectic temperature being 363 °C). This reaction therefore necessitates much lower deposition temperatures (300 °C) than that used for deposition on platinum. This constraint has an advantage in that it renders PSZT deposition on gold compatible with microsystems fabrication processes. Identification of low temperature processing conditions will serve to overcome the aforementioned bottleneck in the realization of devices. Experimental Details
*To whom correspondence should be addressed. E-mail: sharath.sriram@ gmail.com.
Bottom Electrode Deposition. Silicon (100) substrates were dipped in hydrofluoric acid to remove the native oxide layer.
r 2009 American Chemical Society
Published on Web 12/18/2009
pubs.acs.org/crystal
762
Crystal Growth & Design, Vol. 10, No. 2, 2010
Sriram et al.
Table 1. Major Lattice Spacings of PSZT, Gold, and Platinum9,10,12 PSZT orientation (012) (104) (110) (202) (024)
Au d-spacing (A˚) 4.0793 2.9033 2.8661 2.3451 2.0397
orientation
(111) (200)
Pt d-spacing (A˚)
2.3550 2.0390
orientation
(111) (200)
d-spacing (A˚)
2.2650 1.9616
Figure 1. Schematic representation of outcome of PSZT deposition on Au/Ti/Si. The desired bottom electrode configuration was deposited by electron beam evaporation (at room temperature and under vacuum of 1 10-7 Torr). The two bottom electrode configurations used for this study were (i) gold thin films with titanium adhesion layer (Au 150 nm on Ti 15 nm on Si), and (ii) gold thin films with titanium adhesion layer and a silicon dioxide intermediate layer (Au 150 nm on Ti 15 nm on SiO2 200 nm on Si). All bottom electrode layers for each type of sample were sequentially deposited without breaking a vacuum. PSZT Thin Film Deposition. Thin films of PSZT (1.6 μm thick) were deposited on metal-coated silicon substrates by RF magnetron sputtering in a 10% oxygen balance argon atmosphere at a pressure of 10 mTorr from a 100 mm diameter target of composition (Pb0.92Sr0.08)(Zr0.65Ti0.35)O3.11 A 3-in. resistive substrate heater was used, which was compatible with deposition in an oxygen atmosphere, and samples were placed in direct contact with the heater. Very accurate control of temperature was achieved using a model 808 temperature controller programmer (Eurotherm Controls, Inc.). The postdeposition cooling rate was found to influence the degree of perovskite orientation in the thin films,7,8 and so, a cooling rate of 5 °C/min was chosen. Transmission Electron Microscopy Analysis. Cross-sectional transmission electron microscopy (XTEM) specimens were prepared by gluing coated samples face-to-face with backing silicon, using high strength epoxy resin. A core (2.3 mm in diameter) was then ultrasonically machined with the interface of interest at the center of the core. The core was then glued into a brass tube of 3 mm external diameter. This was then sawed into 500 μm thick sections, ground, and ion milled (Arþ at 5 keV) to produce electron transparent cross-sectional specimens. Knowing the silicon wafer normal and by appropriate tilting of the silicon specimen in the transmission electron microscope (TEM), it was possible to ensure that the film was viewed edge-on. TEM analysis was carried out using a JEOL 2010F (200 kV) fitted with a Gatan Imaging filter and an energy dispersive X-ray analysis system. X-ray Diffraction Analysis. X-ray diffraction (XRD) analysis was carried out using a Scintag X-ray diffractometer operating with a cobalt X-ray source (at a wavelength of 0.179020 nm). The scans were carried out for a 2θ range of 20° to 60° with steps of 0.02°. The collected data were shifted to correspond to the copper KR wavelength (0.154056 nm) for comparison with the International Centre for Diffraction Data powder diffraction pattern files available.
Theoretical Basis Deposition of lead zirconate titanate (PZT) compounds has traditionally been carried out on platinized silicon substrates. Platinum deposited at room temperature is nanocrystalline and the substrate deposition temperatures used (600-700 °C) result in a (111) texture. As can be observed in Table 1, this low energy (111) state of platinum has significant lattice mismatch to PZT and PSZT. Considering the oxygenated atmosphere used for deposition of PSZT, inert metals (especially those conforming to standard device fabrication schemes) are desirable. We investigated the possibility of using gold as an alternative to platinum. As Table 1 clearly depicts, there is a collection of significant lattice spacings in common to PSZT and gold. This indicates possible lattice guided growth if PSZT is deposited on gold. Despite this theoretical lattice similarity, gold had not been experimented with for 650 °C depositions previously for two reasons: (i) the gold-silicon eutectic at 363 °C and (ii) the possibility of gold evaporation at lower vapor pressures used for deposition. To overcome these drawbacks, we chose to work toward optimizing deposition of PSZT on gold at 300 °C. The underlying choices were to overcome the aforementioned drawbacks and the expectation of lattice guiding by gold at 300 °C to compensate for the thermal energy driven crystallization at 650 °C. Results and Discussion The gold-coated silicon substrates initially chosen had a thin titanium layer to promote the adhesion of gold to the silicon substrates (the resulting structure was 150 nm gold on 15 nm titanium on silicon). XTEM analysis of samples with PSZT thin films deposited at 300 °C showed evidence of gold reacting with silicon through the titanium adhesion layer to form faceted crystallites projecting into the substrate (Figure 1). The angle between the (100) silicon surface and the bounding planes of the crystallite is 54.7° confirming that
Article
the diffused crystallites are bound by dense (111) silicon planes. The gold-silicon reaction also results in an amorphous silicon dioxide layer on the surface of gold, as a result of the fast outward diffusion by silicon.13 This reaction to form the amorphous oxide is further enabled by the oxygen present in the sputtering gas mixture. The presence of this amorphous layer precludes the gold layer from having a pronounced guiding effect on the subsequently deposited layer of PSZT. Extensive microscopy and diffraction analysis of these thin film samples is presented in refs 14 and 15. To overcome the gold-silicon interdiffusion observed in the initial samples, an intermediate silicon dioxide (SiO2, 200 nm thick) layer was introduced between the metal layers and silicon, and PSZT deposition was carried out at 300 °C. The silicon dioxide layer included between the metal layers and the silicon substrate prevents metal-silicon reactions, and uniform thin film layers can be observed in the cross-sectional transmission electron micrograph (Figure 2a). The resulting PSZT films had a nanocolumnar grain structure. The PSZT deposition temperature initiates crystal growth in the gold layer (which was nanocrystalline when deposited), resulting in significant changes to the texture of gold (Figure 2b). In the absence of gold-silicon interactions, the gold layer is able to reorient so that the most densely packed planes (111) are parallel to the surface, and thus the surface energy is minimized. The formation of the amorphous layer is also
Figure 2. Cross-sectional transmission electron microscopy results: (a) uniform thin film layers can be observed in the cross-sectional transmission electron micrograph and (b) grain growth in gold with preferential (111) texturing (lattice spacing of 0.236 nm).
Crystal Growth & Design, Vol. 10, No. 2, 2010
763
prevented. As a result, the highly textured gold layer was able exert a strong guiding effect on the subsequently deposited PSZT thin films (Figure 3), which developed a pronounced (104) texture (discussed later). This guiding effect results in the (104) PSZT planes being aligned parallel to (111) gold planes (Figure 4); this is a feature at different points along the interface interspersed with regions of random orientation. The ratio of the relative intensities of the (111) and (200) gold X-ray diffraction peaks was ∼100 after PSZT deposition at 300 °C on a sample with the buffer layer of silicon dioxide (Figure 3b); this corresponds to an increase in preferential orientation in the gold layer by a factor of 6 (compared to a sample without the silicon dioxide layer) (Figure 3a). The pronounced (111) orientation in the gold layer improves the preferential orientation of the PSZT thin film;the
Figure 4. The preferentially oriented gold layer appears to have a guiding effect on the PSZT thin film, and this high resolution transmission electron micrograph shows the (104) planes of PSZT (with lattice spacing of 0.305 nm) preferentially aligning parallel to the (111) gold planes.
Figure 3. X-ray diffraction results indicating preferential orientation in the gold and PSZT layers.
764
Crystal Growth & Design, Vol. 10, No. 2, 2010
Sriram et al.
structure simulations presented in Figure 5 verifies the atomic overlap and the ability of gold (111) to exert a guiding effect on PSZT (104). Conclusions
Figure 5. Crystal structure simulation showing the superposition of the gold (111) plane on the PSZT (104) plane. From the overlap of atoms of these planes, it can be concluded that gold can exert a guiding effect on PSZT.
ratio of the X-ray diffraction relative intensities of the two major peaks corresponding to (004) and (104) PSZT orientations increased from 1.57 to 8.52 with the inclusion of the silicon dioxide layer (Figure 3). All PSZT thin film peaks were shifted from expected peak positions, while substrate (gold and silicon) peaks were exactly at expected peak positions. The PSZT rhombohedral unit cell parameters a and c were both about 5.29% larger than expected. This diffraction analysis also showed that the unit cells of the PSZT thin films have a 16% greater volume than theoretically expected.15 Crystal structure simulations carried out using these modified unit cell parameters highlight the small lattice mismatches between the gold (111) and PSZT (104) planes as depicted in Figure 5. From the overlap of atoms of these planes, it can be concluded that gold can exert a guiding effect on PSZT. The simulation for gold was based on a lattice constant a of 0.236 nm and that for PSZT was based on the modified rhombohedral unit cell parameters a of 0.604 nm and c of 1.508 nm. On the basis of Table 1, a guiding effect by gold (111) on PSZT (202) might have been expected. According to the powder diffraction reference,10 the PSZT (104) is five times more intense than the (202) orientation, indicating that the former is the preferred equilibrium state. This is the most likely reason for the observed guiding effect of gold (111) on PSZT (104). It should also be noted that the proximity of the d-spacings of the gold (111) and PSZT (202) peaks thwarts structural comparison using XRD, with no related guiding effects observed in the XTEM results. Importantly, the crystal
A combination of transmission electron microscopy and X-ray diffraction has been used to show that gold exerts a guiding effect on the RF magnetron sputter deposited PSZT thin films. The resulting films have a nanocolumnar grain structure with strong preferential orientation. This is one of the first instances of low temperature deposition of preferentially oriented piezoelectric thin films, with this deposition on silicon making the process favorable to microfabrication processes. This has the potential for enabling incorporation of complex functional oxides into devices with processing at a relatively lower temperature of 300 °C, allowing more versatile device designs and fabrication schemes.
References (1) Voigt, W. Lehrbuch der Kristallphysik; B. G. Teubner: Leipzig and Berlin, 1910. (2) Randall, C. A.; Kim, N.; Kucera, J.-P.; Cao, W; Shrout, T. R. J. Am. Ceram. Soc. 1998, 81, 677. (3) Yu, Y.; Tu, J.; Singh, R. N. J. Am. Ceram. Soc. 2001, 84, 333. (4) Maluf, N.; Williams, K. An Introduction to Microelectromechanical Systems Engineering, 2nd ed.; Artech House: Norwood, 2004; p 29. (5) Jaffe, B.; Cook, W. R.; Jaffe, H. Piezoelectric Ceramics; Academic: New York, 1971; p 136. (6) http://www.ndt-ed.org/GeneralResources/MaterialProperties/ UT/ut_matlprop_piezoelectrics.htm (7) Wasa, K.; Kitabatake, M.; Adachi, H. Thin Film Materials Technology: Sputtering of Compound Materials; Springer-Verlag GmbH & Co. KG: Heidelberg, Germany, 2004. (8) Sriram, S.; Bhaskaran, M.; Holland, A. S. Semicond. Sci. Technol. 2006, 21, 1236. (9) Powder Diffraction Pattern Files, International Centre for Diffraction Data (ICDD, formerly the Joint Committee for Powder Diffraction Studies), Newtown Square, PA 19073, Card 04-0784. (10) Powder Diffraction Pattern Files, International Centre for Diffraction Data (ICDD, formerly the Joint Committee for Powder Diffraction Studies), Newtown Square, PA 19073, Card 04-0025985. (11) Sriram, S.; Bhaskaran, M.; du Plessis, J.; Short, K. T.; Sivan, V. P.; Holland, A. S. Micron 2009, 40, 104. (12) Powder Diffraction Pattern Files, International Centre for Diffraction Data (ICDD, formerly the Joint Committee for Powder Diffraction Studies), Newtown Square, PA 19073, Card 04-0802. (13) Sumida, N.; Ikeda, K. Ultramicroscopy 1991, 39, 313. (14) Sriram, S.; Bhaskaran, M.; Mitchell, D. R. G.; Holland, A. S.; Mitchell, A. Microsc. Microanal. 2009, 15, 30. (15) Bhaskaran, M.; Sriram, S.; Mitchell, D. R. G.; Short, K. T.; Holland, A. S. Thin Solid Films 2008, 516, 8101.
