Vacuum 85 (2010) 55e59
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Fabrication of LSS bottom electrode by PLD M.S. Awan a, *, A.S. Bhatti a, S. Qing b, C.K. Ong b a b
Center for Micro and Nano Devices, Department of Physics, Park Road, Near Tramaree Chowk, COMSATS Institute of Information Technology, Islamabad, Pakistan Center for Superconducting and Magnetic Materials, Department of Physics, National University of Singapore, 117542 Singapore, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 October 2009 Received in revised form 23 March 2010 Accepted 25 March 2010
Polycrystalline LaNiO3/SrTiO3/Si(100) (LSS) conducting substrates were fabricated by pulsed laser deposition (PLD) technique. LSS substrate is a potential candidate for the multiferroic materials for use as bottom electrode. X-ray diffraction (XRD), field emission scanning electron microscopy (FE-SEM) equipped with EDX system, atomic force microscopy (AFM) and electrical resistivity were employed to characterize the films. Buffer layer SrTiO3 (STO) deposited at 700 C resulted in dense, smooth and with crack free features. XRD studies confirmed bi-crystalline [(100), (110)] growth of STO on Si(100) substrate. Deposition of bottom electrode LaNiO3 (LNO) epitaxially followed the buffer layer. EDX analyses determined the chemical composition of the films. The role of oxygen partial pressure during deposition affecting the crystallinity and resistivity of the films was explored in detail. Atomic force microscopy revealed the atomic scale features of the films desirable for functional devices. Resistivity of the conducting film (LNO) was w104 U cm at room temperature. Thus it is demonstrated that LNO/STO/ Si(100) is a suitable conducting substrate for growth of the multiferroic functional materials. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: Polycrystalline Pulsed laser ablation Buffer layer Multiferroic Conductive electrode
1. Introduction In recent years, multiferroic materials have received considerable attention for their potential applications in integrated circuits and functional devices [1]. These applications require fabrication of high quality ferroelectric and magnetic thin films on suitable electrodes, such as metal or conducting oxides. Electrodes are required to have certain properties, such as high metallic conductivity, sufficient resistance against oxidation and good adhesion to the films [2]. Therefore, platinum (Pt), one of the few metals satisfying these requirements, has been used as a bottom electrode in high dielectric constant thin film capacitors [3]. However, Pt as an electrode poses a challenge to obtain a good quality thin film on it due to large lattice mismatch with most of the perovskite oxide materials. Moreover, Pt electrodes often result in the formation of hillocks, which may lead to the degradation of dielectric properties. Also the electrical properties of capacitors (like Pt/PZT/Pt) easily degrade in a hydrogen-containing or plasma environment [4]. As an alternative material, conducting perovskite oxides such as LaSrMnO (LSMO), La0.5Sr0.5CoO3 (LSCO), YBa2Cu3O7 (YBCO), SrRuO3 (SRO) and LaNiO3 (LNO) have been tried as bottom electrode [5e8] for the determination of electrical properties of the multiferroic thin films. Among all these conductive oxides LaNiO3 (LNO) is the most attractive contender for electrode material. This is because of * Corresponding author. Tel.: þ92 34 55112329. E-mail address:
[email protected] (M.S. Awan). 0042-207X/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2010.03.011
its simple crystal structure, uncomplicated synthesis at low temperatures compared to other metal oxides. For multilayers containing LNO, the diffusion of ions into the material (YBCO) during fabrication may be smaller than with the multilayers of other materials. Finally, LNO has only two metal ions in it and its composition is thus easier to reproduce. It has a pseudo-cubic perovskite crystal structure with lattice parameter a ¼ 3.84 A and electrical resistivity <103 U cm at room temperature [9]. LNO has been used as the bottom contact for the study of ferroelectric materials BaSrTiO3 (BST), Pb(ZrTi)O3 (PZT), BaTiO3 (BTO) [10e13]. The (100) and/or (110) oriented perovskite thin films have been shown good multiferroic properties [14]. Thus in order to have good electrical contact, the bottom electrode is preferred to have the same orientation with excellent electrical properties. Significant understanding of LNO has been developed through numerous experimental and theoretical analyses [15e17]. PLD has been applied [18e20] to grow LNO and/or functional materials on variety of substances. However, their growth on silicon substrate is still a challenge due to large lattice mismatch. The complexity of lattice mismatch between LNO and Si substrate is addressed by introducing a buffer layer. This paper focuses on the fabrication of conducting LNO thin film on STO buffered Si(100) substrate by PLD. Insulating behavior and low chemical reactivity of STO makes it ideal candidate for buffer layer. Furthermore it has crystallographic similarity with the multiferroic (BFO) materials. Intentionally (100 and/or 110) phases of STO were preferred for growth on Si substrate to facilitate the epitaxial growth of (LNO and BFO) films.
56
M.S. Awan et al. / Vacuum 85 (2010) 55e59
Here we report the topographic, structural, microstructural and electrical properties of STO and LNO films prepared on (100) orientated silicon substrate. Films were deposited by PLD under different oxygen partial pressures. The LNO/STO/Si(100) would be used as the starting substrate for the growth of Multiferroic thin films at a later stage. To the best of our knowledge this is the first attempt to fabricate a bottom contact LNO/STO/Si(100) for the growth of multiferroic films by PLD technique. 2. Experimental Pulsed laser deposition (PLD) technique was employed to deposit multilayers (LNO and STO) on Si (100) substrate. KrF excimer laser in the energy range 200e250 mJ/pulse was used for ablation of target material. Laser wavelength and frequency were l ¼ 248 nm and f ¼ 3e4 Hz, respectively. The laser beam was focused on the off centered ceramic targets by a quartz lens. During ablation, the target was rotated (30 rpm) in order to reduce nonuniform erosion and to get the homogeneous films. There are few intrinsic restrictions on the target used in a PLD system. A good rule of thumb is that high density and highly homogeneous target yield the best films. A compact and dense target sample is necessary for good quality laser plume and uniform erosion of the target surface. Localized melting due to high energy laser pulses in the porous sample may result in the non-uniform erosion of the target surface. These may create pits and bumps on the laser exposed areas as a result of this plume shape and dimensions may be affected. Ultimately film quality is degraded. The dense and crack free polycrystalline target samples (4 ¼ 15 mm) of SrTiO3 and LaNiO3 were prepared by conventional powder metallurgy technique. Powders were calcined after grinding and before sintering for stress relieving and to develop the desired phase. This also helped to reduce the sintering time and to lower the sintering temperature. The details of preparation technique for ceramic targets are given in the Table 1. In the film growth sequence, STO film as a buffer layer was first deposited and then LNO was deposited as the conductive layer. Silicon substrates were cleaned thoroughly to remove the organic and inorganic residuals from the surface. A three step ultrasonic cleaning process was adopted. The substrates were soaked for two hours in dilute HF solution, acetone and isopropanol with ultrasonic agitation successively. The deposition chamber was evacuated to a base pressure of about (1.3 l05 Pa) using a turbo molecular pump. The deposition was carried out on the heated substrates to get good crystallinity. Before deposition (STO) the substrate was heated (800 C) under vacuum to remove the oxide layer from the substrate surface. To prevent oxide re-growth during the early stages of the growth process, all depositions were initiated at low oxygen partial pressure. After the first (200 laser pulses) the oxygen partial pressure was raised and then maintained at the desired level. In addition to this, during the initial deposition at low oxygen partial pressure, it was expected that depositing particles will take oxygen partially from the substrate surface to initiate the STO phase. Later, STO films were deposited under different oxygen partial pressures ranging from 0.01 Pa to 106.65 Pa at 700 C for 25 min. The selection criteria for oxygen partial pressure was to get
Table 2 Optimized PLD deposition parameters for buffer layer (SrTiO3) and bottom electrode (LaNiO3) thin films. Films
Ts ( C)
ds (cm)
P (Pa)
T (nm)
Td (min)
E (mJ/pulse)
F (Hz)
SrTiO3 LaNiO3
700 600
4 3.5
1.33 39.99
60 437
25 45
250 225
3 4
Ts, substrate temperature; ds, source to substrate distance; P, O2-partial pressure; T, film thickness; Td, deposition time; E, energy of the laser (mJ/pulse); F, frequency/ repetition rate.
the preferred (100) and/or (110) phases for the STO layer. After deposition, STO films were slowly cooled to room temperature under high oxygen pressure in order to maintain the oxygen stoichometry. The optimum deposition conditions for STO buffer layer are given in Table 2. Perovskite LaNiO3 conductive thin films were deposited on STO buffered Si (100) substrate under varied conditions. The optimum conditions for LNO conductive layer are summarized in Table 2. LNO films were deposited under various oxygen partial pressures ranging from 13.33 Pa to 93.33 Pa in order to achieve good electrical conductivity and preferred crystallographic orientations. Post deposition in-situ annealing was carried out at the sintering temperature to avoid oxygen vacancies and maintain the phase stoichometry of the film. This is required to further improve the electrical conductivity of the LNO films [21]. The film was then cooled slowly to room temperature. The phases and structures of STO and LNO thin films were examined by X-ray diffraction with Cu-Ka radiation. The cross-sectional view, surface morphologies, chemical analyses and topography of the films were examined by the FE-SEM and AFM studies. The electrical properties were measured by four-probe technique. 3. Results and discussion The crystallographic phases and orientations of STO buffer layer and LNO conductive thin films deposited on Si (100) were identified by XRD patterns taken at room temperature. The effects of oxygen partial pressure on the phase formation and electrical resistivity of STO and LNO films were studied in detail. Fig. 1 shows the formation of various STO phases as identified in the XRD patterns deposited at 700 C under various oxygen partial pressures. STO films grown at low oxygen partial pressure, i.e., (0.01 Pa) are shown in Fig. 1(a). The peaks are labeled by comparing the XRD pattern with the available literature and standard JCPD card of PDF # 84-
Table 1 Preparation detail of PLD targets for buffer layer (SrTiO3) and bottom electrode (LaNiO3). Target material
Calcination temperature ( C)
Calcination time (min)
Sintering temperature ( C)
Sintering time (min)
Calcining/ sintering atmosphere
SrTiO3 LaNiO3
900 700
60 30
1250 810
180 120
Air Air
Fig. 1. XRD patterns of SrTiO3 (STO) thin films deposited at 700 C on Si(100) substrate by PLD under oxygen partial pressure of (a) 0.01 Pa (b) 1.33 Pa (c) 13.33 Pa and (d) 106.65 Pa.
M.S. Awan et al. / Vacuum 85 (2010) 55e59
57
Fig. 4. XRD patterns of the (a) STO/Si(100) substrate (b) LNO conductive layer deposited on STO/Si(100) substrate at 600 C under 39.99 Pa oxygen partial pressure.
Fig. 2. XRD patterns of (a) Si(100) substrate (b) STO target and (c) STO buffer layer deposited on Si(100) substrate under 1.33 Pa oxygen partial pressure at 700 C.
0444 (ICSD # 201257). Two major peaks (111) and (110) along with the peak from Si substrate were observed. So under low oxygen partial pressure the growth conditions are not favorable for the (100) and/or (110) phases of STO. Desired orientations are absent in the XRD patterns. Contrary to this STO films grown at relatively higher oxygen partial pressures (1.33 Pa), three peaks (100), (110) and (200) were recorded in the XRD patterns as shown in Fig. 1(b). Thus increase in oxygen partial pressure played a key role for the formation of desired phases of STO film. Fig. 1(c) shows the XRD pattern of the STO film deposited at (13.33 Pa) oxygen partial pressure, where (100) peak is now suppressed and an extra peak from (111) plane is reappeared. STO film deposition under rather higher oxygen partial pressure (106.65 Pa) results in the peak pattern as shown in Fig. 1(d). Here (100) peak is again absent and a major unidentified peak at 2q ¼ 44.78 appeared along with some other unwanted peaks like (111) and (211). It is observed that the oxygen partial pressure range is very narrow for the development of preferred STO phases on the silicon substrate. Thus, the STO film deposited at 700 C under (1.33 Pa) oxygen partial pressure was best suited for the buffer layer. Deposition at other pressures promotes undesired phases and orientations which would affect the quality of the interface. The preferential crystallographic orientation of thin film is supposed to minimize the surface and interface energies. Fig. 2 is a comparison of XRD patterns of STO film deposited on Si(100) substrate and bulk ceramic target sample.
Fig. 3. XRD patterns of the LNO thin films deposited at 600 C under various oxygen partial pressures (a) 13.33 Pa (b) 39.99 Pa and (c) 93.32 Pa.
For comparison the XRD pattern of Si(100) substrate is also given. The FWHM of the (110) peak for the STO film deposited under optimum conditions for 25 min is 0.2 . This value is comparable to that of reported by others [22,23]. This shows that the film is well crystallized. This peak is bit shifted as compared to the one usually observed in the bulk samples. The shift is possibly due to the strains developed between the film grains and the substrate surface during the film growth. This is not becoming a big concern here, as we were only interested in the desired (100) and/or (110) phases. The same reason also goes to the shift of other peaks. LNO films on STO/Si(100) substrate were deposited at 600 C for 45 min under various oxygen partial pressures. Their XRD scans are shown in Fig. 3. In the case of LNO films deposited on STO/Si(100) substrate, it was observed that oxygen partial pressure does not significantly affect the structural properties. Variation in the peak intensities confirmed a little effect on the crystallinity of the deposited LNO film under different oxygen partial pressures. The oxygen partial pressure for growth of LNO films was tuned to achieve better electrical properties of these films. The XRD patterns in Fig. 4 show that LNO film epitaxially followed the same peak patterns as was optimized for the STO buffer layer. Only two reflections from the planes (100) and (110) with major reflection coming from (110) plane were recorded for LNO films. Thus it can be concluded that LNO films epitaxially followed the bi-crystalline STO buffer layer on Si(100) substrate. LaNiO3 (LNO) has a perovskite structure like STO and hence under suitable deposition conditions follow the lattice-coherent growth mode. For film thickness measurement, chemical analyses and surface topography, FE-SEM and AFM studies were employed. Fig. 5 shows the cross-sectional view in a FE-SEM micrograph of the STO, LNO
Fig. 5. Field emission scanning electron microscopic(FE-SEM) image showing crosssectional view of the LNO/STO/Si(100) conductive substrate.
58
M.S. Awan et al. / Vacuum 85 (2010) 55e59
Fig. 6. Atomic force microscopic (AFM) images of the (a) STO buffer layer and (b) LNO conductive layer deposited under optimum deposition conditions.
Fig. 7. EDX Spectra showing chemical analyses of (a) STO buffer layer (b) LNO conducting layer. Analyses were performed from top surface of the layers.
and BFO films deposited on Si(100) substrate. It is deduced that STO and LNO films were 60 5 nm and 437 5 nm thick, respectively. The image also shows the top layer of multiferroic (Bi1.1FeO3) film. Fig. 6(a & b) is the (2000 2000) nm2 tapping mode AFM right scans of the STO and LNO film. It is evident that both the films exhibit spherical granular features with smooth, dense and crack free surfaces. The root mean square (Rrms) values of the surface
roughness for the STO and LNO films were 1.50 nm and 2.51 nm respectively. The average grain size of STO and LNO films was 125 10 nm and 200 20 nm respectively. Due to the lattice matching, (LNO) grains heterogeneously nucleate at the smooth surface of the STO grains. As a result of this shape and dimensions of the two types of grains are less or more identical, irrespective of the film thickness. This is shown in AFM images in Fig. 6. LNO
Fig. 8. AFM images (3D) of the (2000 2000) nm2 (a) STO buffer layer and (b) LNO conducting layer.
M.S. Awan et al. / Vacuum 85 (2010) 55e59
59
4. Conclusion
Fig. 9. Electrical resistivity as a function of oxygen partial pressure of the LNO/STO/Si (100) substrate prepared under optimum conditions.
epitaxially followed the optimized buffer layer. This is evident from the XRD graphs in Fig. 4. As the LNO film is smooth and homogeneous therefore it could be used as bottom electrode layer. Chemical analyses of the films were performed by taking EDX spectra from the top surface. Fig. 7(a & b) is the EDX spectra of the STO and LNO films respectively. Fig. 8(a & b) represents the three dimensional (3D) view of the STO and LNO films having atomic scale valley to height distances. These reflect dense, smooth and crack free surfaces with low level of porosity on the surface. These are critical and affect the various film related properties as has been reported earlier [24]. It can also be seen that there are no droplet formation as could be the case in PLD, since the particle sizes are much smaller than those of droplets [25]. The electrical resistivity of the LNO films deposited on STO/Si (100) substrate at different oxygen partial pressures were measured by standard four-probe method. Fig. 9 shows the electrical resistivity of the LNO films as a function of oxygen partial pressure. It is evident that at low oxygen partial pressure (13.33 Pa) the LNO film is highly resistive and then becomes conductive with increase in oxygen partial pressure. From these measurements, LNO films deposited at high oxygen partial pressure, i.e., around (39.99 Pa) were highly conductive (r w 104 U cm). The values of resistivity of these films are as good as obtained by other techniques, such as CSD, sputtering and radio frequency magnetron sputtering [26e28]. These electrical measurements also highlights the key role of oxygen partial pressure to obtain good electrical conducting LNO films on STO buffered silicon substrate. As already discussed, variations in oxygen pressure for LNO deposition strongly affect the electrical properties of the film but this was not observed in XRD patterns. It had already been suggested that stoichometry of the LaNiO3 phase and oxygen vacancies play an important role for conductivity of the LNO films [19,29]. Here we have demonstrated that the use of STO as a buffer layer improves the crystallinity of the LNO film and without modifying the crystal structure, its electrical properties can be tailored by varying the oxygen partial pressures. Thus, it is proposed here that conducting thin films of LNO deposited by using STO as buffer can be produced by controlling oxygen partial pressures during growth. Therefore LNO/STO/Si(100) conducting substrate could be an excellent starting point for fabrication of multiferroic devices. This system also allows the growth of buffer and electrode in a single step.
PLD has proven to be a viable technique for the preparation of multilayer structures of conductive oxides and functional materials. The highly conducting LaNiO3 (LNO) thin films have been successfully fabricated on SrTiO3 (STO) buffered Si(100) substrate by PLD technique. Oxygen partial pressure significantly affected the phases and orientation of the STO films. Optimally deposited STO films exhibited bi-crystalline (100 and 110) growth on Si(100) substrate. This growth mode decreases the interfacial energy by reducing the lattice mismatch between STO and LNO films. The AFM images reveal that the STO and LNO films have atomically smooth surfaces with dense and extremely homogeneous grain distribution, which are highly desirable for an electrode template layer. It is demonstrated that the resistivity of LNO film strongly depends on the oxygen partial pressure during deposition without affecting the crystal structure. The magnitude of electrical resistivity of LNO film deposited on STO/Si(100) substrate under optimized oxygen pressure is suitable to be a bottom electrode for ferroelectric or dielectric materials and can replace traditional metal/oxide electrodes. Acknowledgments The authors thank to Dr. Ma Yunqui and M. Yasar for fruitful discussion and facilitating the AFM studies. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]
Ramesh R. Nature 2009;461:1218. Fang L, Shen M, Yao D. Appl Phys 2005;A80:1763. Lee WJ, Kim YM, Kim HG. Thin Solid Films 1995;269:75. Kim K, Song YJ. Microelectron Reliab 2003;43:385. Ramesh R, Chan WK, et al. Appl Phys Lett 1992;61:1537. Cheung JT, Morgan PED, Neurgaonkar R. In: Proceedings of fourth international symposium on integrated ferroelectrics. Colorado Springs; 1992. p. 518. Wu CM, Hong TJ, Wu TB. J Mater Res 1997;12:2158. Chen MS, Wu JM, Wu TB. Jpn J Appl Phys 1995;34:4870. TaoYu Chen YF, Liu ZG, Xiong SB, Sun L, Chen XY, Ming NB. Mater Lett 1996;26:291. Wu CM, Wu TB. Mater Lett 1997;33:97. Zhu TJ, Lu L, Thompson CV. J Cryst Growth 2004;273:172. Bao DH, Wakiya N, Shinozaki K, Mizutani N, Yao X. Appl Phys Lett 2001;78:3286. Simoes AZ, Ries A, Filho FM, Riccardi CS, Varela JA, Longo E. Appl Phys Lett 2004;85(24):5962. Li AD, Ge CZ, Ming NB. Appl Phys Lett 1996;69:161. Sánchez F, Ferrater C, Alcobé X, Bassas J, Garcıá-Cuenca MV, Varela M. Thin Solid Films 2001;384:200. Kim SB, Maeda T, Yamada Y, Suga T, Yamada Y, Watanabe T, et al. Physica C 2001;979:357. Wakiya N, Azuma T, Shinozaki K, Mizutani N. Thin Solid Films 2002;410:114. Zheng L, Zhu J, Zhang Y, Jiang SW, Li YR, Wei XH, et al. Appl Surf Sci 2006;252:3609. Zhu J, Zheng L, Luo WB, Li YR, Zhang Y. J Phys D Appl Phys 2006;39:2438. Sánchez F, Garcıá-Cuenca MV, Domingo N, Ferrater C, Varela M. Vacuum 2002;64:337. Liu BT, Cheng CS, Li F, Wu DQ, Li XH, Zhao QX, et al. J Alloys Compounds 2007;440:276. Sánchez F, Aguiar R, Trtik V, Guerrero C, Ferrater C, Varela M. J Mater Res 1998;13:1422. Yamada T, Wakiya N, Shinozaki K, Mizutani N. Appl Phys Lett 2003;83:4815. Celinska J, Joshi V, Narayan S, McMillan LC, Araujo P. Appl Phys Lett 2003;82:3937. Lee SG, Hwang DS, Park YK, Park JC. ApplPhys Lett 1994;65:764. Yu S, Yao K, Tay FH. Chem Mater 2004;16:346. Zhu J, Zheng L, Zhang Y, Wei XH, Luo WB, Li YR. Mater Chem Phys 2006;100:451. Guangda HU, Tingao T, Jianbin XU. Jpn J Appl Phys 2002;41:6877. Mizusaki J, Tabuchi J, Matsuura T, Yamauchi S, Fueki KJ. Electrochem Soc 1989;136(7):2082.