This content has been downloaded from IOPscience. Please scroll down to see the full text.
Download details:
IP Address: 140.113.38.11
This content was downloaded on 28/04/2014 at 05:29
Please note that terms and conditions apply.
The Effects of Low-Pressure Rapid Thermal Post-Annealing on the Properties of (Ba,
Sr)TiO3 Thin Films Deposited by Liquid Source Misted Chemical Deposition
View the table of contents for this issue, or go to the journal homepage for more 2001 Jpn. J. Appl. Phys. 40 L1333
(http://iopscience.iop.org/1347-4065/40/12A/L1333)
Jpn. J. Appl. Phys. Vol. 40 (2001) pp.L1333–L1335 Part 2, No. 12A, 1 December 2001
c
2001 The Japan Society of Applied Physics
The Effects of Low-Pressure Rapid Thermal Post-Annealing on the Properties
of (Ba, Sr)TiO
3Thin Films Deposited by Liquid Source Misted Chemical Deposition
Ming-Jui YANG, Chao-Hsin CHIEN, Ching-Chich LEU, Ren-Jian ZHANG1, Shich-Chuan WU, Tiao-Yuan HUANGand Tseung-Yuen TSENG1
National Nano Device Laboratories, Hsinchu 30050, Taiwan, R.O.C.
1Institute of Electronics, National Chiao Tung University, Hsinchu 30050, Taiwan, R.O.C.
(Received August 27, 2001; accepted for publication October 11, 2001)
The technique of low-pressure post-annealing process with additional second-step annealing for preparation of the Ba0.7Sr0.3TiO3thin films deposited by liquid source misted chemical deposition (LSMCD) has been proposed. With employing this annealing procedure, the leakage current density can be significantly eliminated by approximately one order of magnitude at 2 V. In particular, process temperature can be reduced from 650◦C to 600◦C without suffering deteriorated crystallinity issue, which is identified by both C–V measurement and X-ray diffraction spectrum. The extracted dielectric constant is 310 with extreamly low loss tangent of 0.005. The spectrum of atomic force microscopy (AFM) shows that this low-pressure process results in smoother surface topography. Moreover, thermal desorption spectrums assure that less residual organics and contam-inations were left after low pressure post-annealing. This may be one of the reasons for lowering crystallization temperature and the improved electrical properties.
KEYWORDS: (Ba, Sr)TiO3, LSMCD, low pressure rapid thermal post-annealing, process temperature, leakage current density, residue organics
The projected capacitor area in the dynamic random ac-cess memory (DRAM) cell is rapidly shrinking as the inte-gration density increases. The required capacitance for as-suring stable circuit operation and sufficient soft-error immu-nity, which is at 25 fF/cell, could be achieved either by reduc-ing the thickness of the dielectric films or by increasreduc-ing the effective surface area of the capacitor electrodes. However, neither approach seems to be practical to next-generation of ULSI DRAMs fabrication because of the unacceptable leak-age current increase and the extreamely complicated integra-tion process, respectively. Therefore, the incorporaintegra-tion of new dielectric material with high dielectric constant becomes the most likely alternative approach for the coming giga-bit dynamic random access memories. Recently, (Ba, Sr)TiO3
(BST) with excellent dielectric properties, such as large di-electric constant, low didi-electric dissipation factor, and low leakage current level, has attracted extensive attention for the application in next generation ULSI DRAMs.1, 2)
Al-though many techniques have succeeded in depositing high-quality BST thin films, for example, rf-magnetron sputtering, pulsed laser ablation, and metal-organic chemical vapor depo-sition (MOCVD),3–5) liquid source misted chemical deposi-tion (LSMCD), in particular, has been proved to be an simple and inexpensive alternative technique with the capability of precise composition control and excellent step coverage.6, 7) In this paper, we study the effects of low-pressure rapid ther-mal post-annealing on the properties of BST thin films de-posited by LSMCD. It has been found that low-pressure rapid thermal post-annealing can lower down the BST film crystal-lization temperature by 50◦C and, moreover, reduce leakage current by near one order of magnitude at 2 V.
The Ba0.7Sr0.3TiO3 films prepared by LSMCD method
were deposited on Pt/SiO2/Si substrates. The as-deposited
samples were subjected to drying at 150◦C to remove the residual solvent and then pyrolysis at 400◦C for 10 min. The desired BST film thickness of about 100 nm was achieved through multiple repetitions of the process sequence. Sub-sequently, some samples accepted rapid thermal annealing
in atmospheric oxygen ambient at 650◦C for 1 min and oth-ers, for comparison, were performed with low-pressure oxy-gen rapid thermal annealing at 50 Torr and 600◦C for 10 min. For electrical measurements, Pt top electrodes were deposited onto BST films through a shadow mask with a diameter of 0.2 mm. After the top electrode formation, additional second-step annealing at 600◦C for 30 min in air was selectively ap-plied to the samples. The electrical properties and reliabil-ity characteristics of the metal oxide were measured by using a Hewlett-Packard (HP) 4156 semiconductor parameter ana-lyzer. X-ray diffraction (XRD) was also used to identify the composition and the phase of these new metal oxide films. Surface morphologies of BST thin films were characterized by atomic force microscope (AFM). The residual gases and contamination left inside the bulk of thin films were detected by thermal desorption spectrums (TDS).
The electrical properties of deposited BST thin films are shown in Fig. 1. Figure 1(a) exhibits the leakage current den-sities and Fig. 1(b) shows both the dielectric constants and dissipation factors of BST thin films as a function of applied top gate bias for various annealing conditions. In Fig. 1(a), the sample annealed at 760 Torr and 650◦C for 1 min shows a slightly higher leakage current than that with low pressure post-annealing at lower temperature in the low field regime. However, there is no obvious difference above the knee be-tween these two samples. The leakage current densities are around 3.1E-7 A/cm2 at 2 V. On the other hand, the
leak-age current density of the sample subjected to low-pressure annealing is found tremendously suppressed with additional second-step annealing at 600◦C in air for both relaxation and Schottky emission currents. The current density is about 4.45×10−8A/cm2at 2 V, which is almost one order of
magni-tude improvement compared to the previous samples. The im-pacts of low-pressure process on the BST thin film are further identified by the film capacitance behavior, as shown in Fig. 1(b). The trend is very different from that of leakage current. The capacitance decreases dramatically for the sample treated with only one-step low-pressure post-annealing. As we added
L1334 Jpn. J. Appl. Phys. Vol. 40 (2001) Pt. 2, No. 12A M.-J. YANGet al.
Fig. 1. (a) Leakage current densities, (b) dielectric constants and dissipa-tion factors as a funcdissipa-tion of the applied bias voltage for the samples various annealing conditions.
second-step annealing, the film capacitance is almost recov-ered and dissipation factor becomes even ramarkably smaller. The capacitance recovery is speculated to the microstructure transformation of BST films and will be discussed later. The dissipation factor is believed to closely related to the leakage current density, thus the films subjected to 2-step low-pressure annealing showed lower dielectric loss.
In order to obtain the picture of the detailed mechanism about the film improvement by low-pressure annealing pro-cedure, the impurities contained in the bulk, surface morphol-ogy and microstructure of BST thin films were analyzed with thermal desorption spectrum, atomic force microscope and X-ray diffraction spectrum, respectively. Figure 2 shows the level of residual organics and contamination, including CO2,
H2O and O2, left in the BST films after high- and low-pressure
annealing conditions by TDS. It is clearly seen that low-pressure annealing can more efficiently remove residual or-ganics and contamination. Figure 3 shows the surface mor-phologies of the BST thin films with various heat treatment sequences. Both thin films subjected to low-pressure anneal-ing, shown in Figs. 3(a) and 3(b), exhibit much smoother surface morphologies than that of thin film with 1 atm post-annealing. The values of surface roughness are seperately
Fig. 2. TDS spectra of BST thin films after (a) 760 Torr for 1 min at 650◦C (b) 50 Torr for 10 min at 600◦C.
Fig. 3. AFM images of BST thin films after (a) 50 Torr for 10 min at 600◦C (b) 50 Torr for 10 min and the second-step annealing for 30 min at 600◦C (c) 760 Torr for 1 min at 650◦C.
Jpn. J. Appl. Phys. Vol. 40 (2001) Pt. 2, No. 12A M.-J. YANGet al. L1335
Fig. 4. XRD patterns of BST thin films after (a) 760 Torr for 1 min. at 650◦C (b) 50 Torr for 10 min. at 600◦C (c) 50 Torr for 10 min and the second-step annealing for 30 min at 600◦C.
1.839 nm, 2.122 nm, and 2.738 nm. However, without second step annealing, abundant porosities have been observed in the film. The addition of another annealing in air at the same tem-perature has been found to successfully eliminate the number of porosity and make the film structure much denser. Ac-cording to the previous reports the leakage current consists of two components, i.e., relaxation current and Schottky emis-sion current. It is believed that relaxation current is strongly correlated to the concentration of impurities and oxygen va-cancies in the bulk, while Schottky emission current is mainly dictated by the film/electrode interface quality.8–11)From our
leakage current results, we speculate that the variations due to different heat treatments are ascribed to the plenty of oxygen vacancies produced in thin films during the lower-pressure annealing step. Although low-pressure annealing process is beneficial for removing the residual organics and contamina-tion contained in thin films,12–14)it would also induce more oxygen atoms evaporated at the same time and then create oxygen vacancies distributed in thin films during crystalliza-tion. While smoother surface has been obtained, those gener-ated oxygen vacancies can deteriorate the film crystalline mi-crostructure and make them more porous. Consequently, no improvement has been achieved in the film J –V characteris-tic. Therefore, our intention of employing additional second-step annealing is trying to provide the required oxygen and energy to recovery vacancies.14)The results do show that
ad-ditional annealing can make thin film denser without causing significant increase of surface roughness, which lead to re-duction of both relaxation and Schottky emission currents.
Figure 4 shows XRD patterns of those corresponding BST thin films. From these spectra, we try to find out the ori-gin of the capacitance behaviors induced by low-pressure
an-nealing. As we can see, the thin film with 760 Torr, 650◦C post-annealing exhibits a strong (110) texture. In contrast, low-pressure annealing favors the preferential growth of (111) texture and would be converted into (110) oriented texture by the additional post-annealing in air at 600◦C. Consistent with previous reports, the development of stronger (110)-oriented BST thin film does show better crystallinity and higher dielec-tric constant.15–17)Thus, we believe that these preferred ori-entations of the films with distinct heat treatments contribute to the variations of the film capacitance behavior.
In summary, a new low-pressure processing with the second-step annealing technique for preparation of the Ba0.7Sr0.3TiO3thin films deposited by LSMCD has been
pro-posed. Not only lower crystallization temperature but also more excellent electrical properties can be obtained by this newly developed method. The improvement is believed due to the less remain of residual organics and contamination in the BST thin film, the smoother and denser film structure and reduction of oxygen vacancies. Therefore, this method is very useful for achieving highly integrated memory applications.
1) Y. Ohno, T. Horikawa, H. Shinkawata, K. Kashihara, T. Kuroiwa, T. Okudaira, Y. Hashizume, K. Fukumoto, T. Eimori, T. Shibano, K. Arimoto, H. Itoh, T. Nishimura and H. Miyoshi: Proc. Symp. VLSI
Tech-nology, 1994, p. 149.
2) K. Takemura, T. Sakuma and Y. Miyasaka: Appl. Phys. Lett. 64 (1994) 2967.
3) C. M. Chu and P. Lin: Appl. Phys. Lett. 70 (1997) 249. 4) S. Yoon and A. Safari: J. Appl Phys. 76 (1994) 2999.
5) C. S. Chern, S. Liang, Z. Q. Shi, S. Yoon, A. Safari, P. Lu, B. H. Kear, B. H. Goodreau, T. J. Marks and S. Y. Hou: Appl. Phys. Lett. 64 (1994) 3181.
6) M. Huffman: Integr. Ferroelectr. 10 (1995) 39.
7) H. J. Chung, J. H. Kim, W. S. Moon, S. B. Park, C. S. Hwang, M. Y. Lee and S. I. Woo: Integr. Ferroelectr. 12 (1996) 185.
8) T. Horikawa, T. Makita, T. Kuroiwa and N. Mikami: Jpn. J. Appl. Phys. 34 (1995) 5478.
9) S. O. Park, C. S. Hwang, H. J. Cho, C. S. Kang, H. K. Kang, S. I. Lee and M. Y. Lee: Jpn. J. Appl. Phys. 35 (1996) 1548.
10) Y. Fukuda, H. Haneda, I. Sakaguchi, K. Numata, K. Aoki and A. Nishimura: Jpn. J. Appl. Phys. 36 (1997) L1514.
11) Y. Fukuda, K. Numata, K. Aoki, A. Nishimura, G. Fujihashi, S. Okamura, S. Ando and T. Tsukamoto: Jpn. J. Appl. Phys. 37 (1998) L453.
12) Y. Ito, M. Ushikubo, S. Yokoyama, H. Matsunaga, T. Atsuki, T. Yonezawa and K. Ogi: Jpn. J. Appl. Phys. 35 (1996) 4925.
13) N. Ogata, M. Nagata, K. Ishihara, H. Urashima, A. Okutoh, S. Yamazaki, S. Mitarai and J. Kudo: Jpn. J. Appl. Phys. 37 (1998) 3481. 14) Y. Fujimori, T. Nakamura and H. Takasu: Jpn. J. Appl. Phys. 38 (1999)
5346.
15) T. Horikawa, N. Mikami, T. Makita, J. Tanimura, M. Kataoka, K. Sato and M. Nunoshita: Jpn. J. Appl. Phys. 32 (1993).
16) S. Y. Cha, H. C. Lee, W. J. Lee and H. G. Kim: Jpn. J. Appl. Phys. 34 (1995) 5220.
17) J.-G. Cheng, X.-J. Meng, B. Li, J. Tang, S.-L. Guo and J.-H. Chu: Appl. Phys. Lett. 75 (1999) 2132.
