• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Characterization of Physical and Electrical Properties of BaTiO^Sub 3^ Films Deposited on P-Si By Modified Polymeric Precursors

Posted on: Sunday, 18 September 2005, 03:01 CDT

Highly uniform BaTiO^sub 3^ (BTO) films with thickness well below 100 nm were deposited on p-Si by spin coating using a modified polymeric precursor method. The polymeric precursor gel was redissolved into glacial acetic acid to improve the wetting property of the spinning solution to the Si substrates (2.5-in. diameter). The morphology, composition, thickness, and refractive index of the films were investigated using x-ray diffraction (XRD), scanning electron microscopy (SEM), energy-dispersive x-ray spectroscopy (EDS), and ellipsometry. The films are found to be polycrystalline. They exhibit uniformity over the whole wafer in regard to thickness, composition, and absence of surface features. The capacitors constructed with the BTO films on Si were further investigated by electrical characterizations. Current-voltage (I-V) measurements reveal a leakage current due to a Poole-Frenkel mechanism. The energy gap is evaluated to be 3.95 eV. A metal-insulator- semiconductor (MIS) behavior is observed through capacitance- conductance-voltage (C-G-V) measurements. The interface state density (D^sub it^) at the BTO/p-Si interface is estimated to be of the order of 10^sup 12^ eV^sup -1^ cm^sup -2^.

Key words: High-k dielectrics, metal-oxide-semiconductor (MOS) capacitors, electrical characterization, interface states, polymeric precursor method, barium titanate

INTRODUCTION

Alternative high-k dielectrics to SiO^sub 2^ have recently attracted great attention for applications in electronic devices.1- 3 One promising class of materials is ferroelectrics, such as SrTiO^sub 3^ (STO), BaTiO^sub 3^ (BTO), and (Ba^sub 1-x^Sr^sub x^)TiO^sub 3^ (BST).1,3,4

Barium titanate (BaTiO^sub 3^) and the other perovskite-type materials are being extensively studied for their potential commercial applications in metaloxide-semiconductor (MOS) capacitors due to their desirable dielectric properties.3-8 Many methods, such as sputtering,9 laser ablation,10 metalorganic chemical vapor deposition (MOCVD),11 and sol-gel12 techniques, have been used to grow thin films; however, most of these methods are expensive and require complex equipment. The polymeric precursor method13-16 is a chemical technique that offers several advantages, such as low cost, good compositional homogeneity, high purity, relatively low processing temperatures, and the ability to coat large substrate areas. In this technique, the desired metal cations are chelated in a solution using a hydroxycarboxylic acid as the chelating agent. The solution is mixed with a polyhydroxyalcohol and heated to promote esterification reactions in the solution, while the metals remain homogeneously distributed in the polymeric network. However, we found it problematic to obtain uniform films by spin coating the aqueous polymeric precursor solution directly on Si, probably due to the poor wettability of the aqueous solution to the hydrophobic Si surface. The problem of uniformity can be alleviated by using concentrated aqueous solution, but this will typically result in films with thickness well above 100 nm by only one spin of the solution, which is undesirable for use as dielectrics in MOS devices. In this study, we have modified the polymeric precursor method by redissolving the polymeric precursor gel into glacial acetic acid, and have for the first time, to the best of our knowledge, succeeded in obtaining highly uniform BTO films directly on Si substrates with thickness well below 100 nm.

Metal-oxide-semiconductor capacitors (MOSCAP) with BST, BTO, and STO films as the dielectric, as in metal-insulator-semiconductor (MIS) structures, have been reported and analysed.8,17,18 In all cases, reliability issues have risen due to process related defects, leakage currents, or abnormal behavior of the dielectric. These issues have been reported and analysed concerning different materials and deposition techniques while recent reports indicate that degradation mechanisms need to be addressed in more detail if more complex systems are to be built.6'1819 In this paper, we will report the electrical properties of MOSCAPs constructed with the BTO films deposited on p-Si using our modified polymeric precursor method. The aim is to study the electrical properties, identify possible defects, and compare the results to those recently published on similar capacitors built with sputtered films deposited at low temperature.17,20 Issues such as the leakage current and its origin, as well as the capacitive and conductive response of the MOSCAPs are examined.21

EXPERIMENTAL PROCEDURE

The BTO films were prepared by a modified polymeric precursor method. Initially, Ti (IV) isopropoxide was added into a citric acid solution (60-70C) to form titanium citrate. After homogenization of the Ti-citrate solution, a stoichiometric amount of BaCO^sub 3^ was slowly added under stirring. Ammonium hydroxide was also added to promote the complete dissolution of BaCO3 (the final pH of the solution was 7-8). After homogenization of the solution, ethylene glycol was added to promote citrate polymerization by polyesterification reactions. The molar ratio of Ba/Ti was 1:1, the citric acid/Ba molar ratio was fixed at 1.00, and the citric acid/ ethylene glycol ratio was fixed at 60/40 (mass ratio). Polymerization was performed at 9O0C for 12 h in an open-mouthed 500 mL beaker (evaporated water was compensated for periodically). At the final stage, more water was not added so that the polymeric precursor solution became a highly viscous gel. A solution for spin coating was prepared by redissolving 0.25 g of the gel with 10 mL of glacial acetic acid. Polished Si (100) wafers (p-type, 2.5 in.) were freshly treated with HF solution before use as substrates. The solution was deposited onto the Si substrates by spinning at 3000 rpm for a period of 30 sec. An atmosphere of acetic acid was created over the spinning substrate by forming a pool of acetic acid underneath the spinning stage and covering up the system with an inverted cup. After deposition, the films were dried on a hot plate at ~150C for 2 min to remove residual solvents. Heat treatment was carried out in a furnace under air at atmospheric pressure with a preset temperature and all samples were treated at 450C for 2 h.

The thickness and refractive index of the BTO films on Si were measured with an ellipsometer (Rudolph Research Auto-El III, at λ = 632.8 nm) and were averaged over 3-5 different measuring points.

X-ray diffraction spectra were recorded using a Philips Goniometer PW1050-25, with the CuK^sub a^ line.

Analysis with scanning electron microscopy (SEM) and energy- dispersive x-ray spectroscopy (EDS) was carried out on a JSM-6300 scanning electron microscope equipped with the EDS detector 6699 (window ATW2). The atomic analysis was performed with the program Link-ISIS.

The resistivity of the Si substrates was in the range of 100-500 ohm cm. Construction of effective electronic devices was made possible by the evaporative deposition of Al metal electrodes on the back of the Si substrate in order to form an ohmic contact and on top of the BTO films via a mask allowing the creation of 1-mm, 2- mm, and 3-mm dots. Thus, devices with the structure of Al/BTO/p-Si were prepared.

For the electrical measurements, the samples were kept in a cryostat under a controlled reduced He atmosphere. The temperature was controlled using an Oxford ITC503 Temperature Controller with a resolution of 0.01 K.

The current-voltage (I-V) curves were taken using a Keithley 617 programmable electrometer connected with an Oltronix power supply, in the temperature range from 213 K to 333 K.

The dielectric response measurements, C-V, G-V, and admittance spectroscopy &969; were performed using a Novocontrol Alpha- N Dielectric Response Analyzer suitably controlled by the WinData software package at room temperature.

RESULTS AND DISCUSSION

The BTO films on Si deposited with our modified polymeric precursor method look to the naked eyes very uniform over a whole silicon wafer (2.5-in. diameter), free of striations and interference color. This is in agreement with the ellipsometry measurements, which show that the film thickness and refractive index are consistently 57.1 nm (0.3 nm) and 1.802 (0.003), respectively, measured on 3-5 points well separated over the whole film surface.

The crystallinity of the BTO thin films was investigated by x- ray diffraction (XRD). Figure 1 depicts the XRD spectrum of a 570 A film. Three peaks are present, which correspond to the BTO cubic structure. An additional peak present corresponds to the Al contacts used as electrodes. Thus, the thin films under investigation are of a polycrystalline nature.

Fig. 1. X-ray diffraction spectrum of a 570 BTO thin film.

Fig. 3. I-V characteristics under forward and reverse bias conditions.

The morphology and composition of the BTO films on Si were investigated with a scanning electron microscope combined with EDS capability. From SEM analysis (not shown here), no topological features are present, confirming that the film surface is continuous and uniform. Figure 2 presents an EDS spectrum of BTO thin films. Ti \Ka1, Ti Kb1, Ba La1, Ba Lb1, and O Ka1 are present at energies 4.511 KeV, 4.932 KeV, 4.466 KeV, 4.828 KeV, and 0.525 KeV, respectively. The peak from the Si substrate is also present. The atomic analysis over more than 20 areas on the thin films revealed an atomic stoichiometry of Ba, Ti, and O to be 20.58%, 19.62%, and 59.80%, respectively. These values confirm a very good quality of the BTO thin films.

The above results show that highly uniform thin films of BTO on Si can be achieved with our modified polymeric precursor method. In contrast, spinning the aqueous polymeric precursor solution typically results in uncontinuous films on the Si substrates. This suggests that redissolving the polymeric precursor gel into glacial acetic acid is effective in largely improving the wettability of the spinning solution to the hydrophobic Si surface. On the other hand, the atmosphere of acetic acid kept over the spinning substrate during the spinning process is thought to have contributed to the uniformity and ultrathinness of our BTO films. Supporting our point of view, a saturated ethanol atmosphere created over the spinning substrate has been shown by Du et al.22 to result in striation-free films with reduced thickness.

Fig. 2. EDS spectrum of BTO thin film.

Fig. 4. Plot of (I/V) versus V^sup 1/2^ as a function of temperature.

The leakage current through the BTO dielectric was measured and the I-V curve is depicted in Fig. 3 for forward and reverse bias. It is evident that both directions show an equal amount of leakage current, proving the presence of a leaky insulator film. The question that must be addressed at this point is the origin of the leakage current. Literature reports, dating back 10 years, refer to such leakage currents for a variety of devices with high-k dielectrics.2,6,8,18,19 The origin of these currents was primarily attributed to either space charge limited currents or PooleFrenkel related effects.1,6,8 In order to determine the mechanism which causes the leakage current, several conduction mechanisms were tested. The best results of these tests were given by the PooleFrenkel mechanism and they are presented in Fig. 4.

The magnitude of the leakage current varies from 10^sup -9^ A to 10^sup -6^ A, depending on the applied voltage. Although these values are rather high, the films investigated can still be considered good insulators, and the effects of this leakage current on the device characteristics can be counted for. All measurements reported hereafter have been corrected for d.c. leakage current.

In order to identify the accumulation-depletion-inversion regions of the device, the capacitance-voltage (C-V) and conductance- voltage (G-V) characteristics for various frequencies were determined simultaneously. Furthermore, in order to account for a dispersion observed in the accumulation region, the data were treated according to the method proposed by Vogel et al.18 All measurements presented hereafter have been treated accordingly.

Figure 6 depicts the C-V characteristics of the Al/BTO/p-Si devices for two different frequencies, namely, 1 KHz and 100 KHz. Extra care needs to be taken for the capacitive response according to the method proposed by Kwa et al.23 The overall behavior is indeed that of a MOSCAP, with the three distinct regimes of accumulation-depletion-inversion clearly shown.

Prior to data analysis, the measurements were corrected for series resistance using a well-defined method developed by Nicollian and Brews.21 Corrections due to the series resistance of the back contact were also made.

Although the films investigated are thick for very large scale integration (VLSI) applications, it is useful to evaluate the density of interface states in order to calculate the parasitic capacitance due to the charges located at the BTO/Si interface. Admittance spectroscopy can identify both bulk and interface electrical defects. For this purpose, the devices have to be properly biased in depletion and a superimposed a.c. signal should be applied across the diode terminals.21 Then, using the parallel conductance versus frequency (Gp/ω versus log ω) for certain d.c. biases in depletion, the value of the density of the interface states, and therefore the parasitic capacitance, can be calculated. Admittance spectroscopy data were collected and analyzed. In Fig. 7, the G^sub p^/ω versus log ω curves for three different biases in depletion are presented. The calculated values of the D^sub it^ obtained from the statistical model21 were found to range between 2 x 10^sup 12^ eV^sup -1^ cm^sup -2^ and 1.8 x 10^sup 13^ eV^sup -1^ cm^sup -2^.

Fig. 5. Plot of term B given in Eq. 2 as a function of 1/T.

Fig. 6. C-V curves corrected for leakage current and series resistance, for two frequencies: ([white circle]) 1 kHz and ([black square]) 100kHz. The three distinct regimes of accumulation, depletion, and inversion are shown.

Fig. 7. The Gp/ω curves for three biases in depletion: -1 V ([white circle]), -0.9 V([black square]), and -0.8V(Δ).

The value of the dielectric constant can be calculated to be equal to 60. This calculation is made using the high frequency curve of 100 kHz and the corrected capacitance value at accumulation. Since the thickness of the film is uniform as shown by the structural characteristics, the measurement of the dielectric constant can be considered accurate enough for the thickness resolution we obtain.

CONCLUSIONS

We have succeeded in depositing thin films of BTO directly on Si by using a modified polymeric precursor method. The XRD data revealed the polycrystalline nature of the thin films. The SEM analysis confirms that the BTO films exhibit a smooth and uniform surface. The EDS analysis confirms the atomic ratio of Ba/Ti/0 in the films to be 1:1:3. So we conclude that our modified wet chemical technique is capable of producing high quality BTO films.

The electrical properties of the capacitors fabricated, regarding the leakage currents and the density of the interface states were identified and analyzed. The leakage currents follow a bulk dominated Poole-Frenkel mechanism. This mechanism is better controlled in devices than the space charge limited (SCR) currents observed in sputtered films. This is mainly due to the nature of the PooleFrenkel emission mechanism, where the behavior of the dielectric is the main reason behind this leakage current, while in the case of SCR parameters such as geometry and operating voltages may affect the current value. The energy gap was evaluated to be 3.95 eV, from I-V measurements as a function of temperature. The density of interface states (Dit), for the BTO/p-Si system deposited by the wet chemical technique, was evaluated to be in the order of 10^sup 12^ eV^sup -1^ cm^sup -2^. Moreover, the dielectric constant of 60 is substantially higher than the value of 16 that has been reported for films deposited by similar technique.24 Therefore, it is suggested that our modified polymeric precursor method can create BTO films suitable for use in various electronic applications, similar to those constructed with sputtering, gaining in lower cost and better control over the stoichiometry.

Future work involves the study of the temperature effect on the device behavior such as post thermal annealing, variable deposition temperatures, and temperature dependent measurements. Also a future target is the preparation of uniform films on Si much thinner than currently achieved in this work by optimizing the processing parameters of our modified wet chemical technique. These procedures are currently in progress.

ACKNOWLEDGEMENTS

The authors thank Drs. D. Anastassopoulos and A. Vradis, Physics Department, University of Patras, for providing the XRD spectra.

(Received February 18, 2005; accepted May 26, 2005)

REFERENCES

1. A.I. Kingon, J.P. Maria, and S.K. Streiffer, Nature (London) 406, 1032 (2000).

2. S. Ezhilvalavan and T.Y. Tseng, Mater. Phys. Chem. 65, 227 (2000).

3. K. Kim, C.G. Hwang, and J.G. Lee, IEEE Trans. Electron. Dev. 45, 598 (1998).

4. S.K. Dey and J.J. Lee, IEEE Trans. Electron. Dev. 39, 1607 (1992).

5. E. Evangelou, N. Konofaos, and C.B. Thomas, Phil. Mag. B 80, 395 (2000).

6. R.A. McKee, F.J. Walker, and M.F. Chisholm, Science 293 468 (2001).

7. RC. Joshi and S.B. Krupanidhi, J. Appl. Phys. 73, 7627 (1993).

8. S. Yamamichi, A. Yamamichi, D.G. Park, T.J. King, and C.M. Hu, IEEE Trans. Electron. Dev. 46, 342 (1999).

9. K. lijima, T. Terashima, K. Yamamoto, K. Hirata, and Y. Bando, Appl. Phys. Lett. 56, 527 (1990).

10. M.-B. Lee, M. Kawasaki, M. Yoshimoto, and H. Koinuma, Appl. Phys. Lett. 66, 1331 (1995).

11. L.A. Wills, B.W. Wessels, O.S. Richeson, and T.J. Marks, Appl. Phys. Lett. 60, 41 (1992).

12. T. Hayashi, N. Oji, and H. Maiwa, Jpn. J. Appl. Phys., Part 1 33,5277(1994).

13. V. Agarwal and M.L. Liu, J. Mater. Sd. 32, 619 (1997).

14. RM. Pontes, E.B. Araujo, E.R. Leite, J.A. Eiras, E. Longo, J.A. Varela, and M.A. Pereira-da-Silva, J. Mater. Res. 15, 1176(2000).

15. E.J.H. Lee, F.M. Pontes, E.R. Leite, E. Longo, J.A. Varela, E.B. Araujo, and J.A. Eiras, J. Mater. Sd. Lett. 19, 1457 (2000).

16. EM. Pontes, E.J.H. Lee, E.R. Leite, E. Longo, and J.A. Varela, J. Mater. Sd. 35, 4783 (2000).

17. Z. Wang, V. Kugler, U. Helmersson, N. Konofaos, E.K. Evangelou, S. Nakao, and P. Jin, Appl. Phys. Lett. 79, 1513 (2001).

18. E.M. Vogel, W. Kirklen Henson, C.A. Richter, and J.S. Suehle, IEEE Trans. Electron. Dev. 47, 601 (2000).

19. A. seekamp, A. Avellan, S. Schwantes, and W. Krautschneider, Int. J. Electron. 90, 607 (2003).

20. Z. Wang, V. Kugler, U. Helmersson, N. Konofaos, E.K. Evangelou, S. Nakao, and P. Jin, Phil. Mag. B 82, 891 (2002).

21. E.H. Nicollian and J.R. Brews, MOS (Metal-Oxide- Semiconductor) Physics and Technology (New York: Wiley, 1982).

22. X.M. Du, X. Orignac, and R.M. Almeida, J. Am. Ceram. Soc. 78,2254(1995).

23. K.S.K. Kwa, S. Chattopadhyay, N.D. Jankovic, S.H. Ols\en, L.S. Driscoll, and A.G. O'Neill, Semicond. Sd. Technol. 18, 82 (2003).

24. R. Thomas, D.C. Dube, M.N. Kamalasanan, and N. Deepak Kumar, J. Sol.-Gel Sd. Technol. 16, 101 (1999).

N. KONOFAOS,1,5 ZHONGCHUN WANG,2 S.N. GEORGA,3 C.A. KRONTIRAS,3 M.N. PISANIAS,3 J. SOTIROPOULOS,3 and E.K. EVANGELOU4

1.-Computer Engineering & Informatics Department, University of Fatras, GR-26500 Patras, Greece. 2.-Department of Chemical and Nuclear Engineering, University of New Mexico, Albuquerque, NM 87106. 3.-Department of Physics, University of Patras, GR-26500 Patras, Greece. 4.-Department of Physics, University of Ioannina, 45110 Ioannina, Greece. 5.-E-mail: nkonofao@ceid.upatras.gr

Copyright Minerals, Metals & Materials Society Sep 2005

Source: Journal of Electronic Materials

More News in this Category