2-1 Low Dielectric Constant Materials
As integrated circuit dimensions continue to shrink, interconnected RC (resistance × capacitance) delay becomes an increasingly serious problem. In order to overcome the RC time delay, incorporating new materials of low resistivity and low permittivity into interconnected structure is one solution to replace the traditional Al and SiO2. Copper is a substitute for aluminum in the future interconnection fabrication1. However, we do not know what kind of low dielectric constant (low k) material will replace the traditional SiO2 or SiOF2, 3, when a dielectric film requires a dielectric constant k < 3.0. The future low dielectric constant materials which support the semiconductor are illustrated by SIA roadmap (Fig. 2-1).
Figure 2-1 summarizes the timeline of the materials anticipated in future generation multi-level interconnected structures. (Source: The International Technology Roadmap for Semiconductors: Update 2001)
Table 2-1 describes the low dielectric constant materials for ULSI interconnection, and the basic requirements for low k materials are summarized in Table 2-2. The hydrogen silsesquioxane (HSQ) film and Methyl-silsesquioxane (MSQ) film which belong to SiOF base films were used in current 130 nm generation of semiconductor industry. So far maybe the k = 2.7 of black diamond film which belongs to SiOCH base film will be used in 90 nm generation of semiconductor industry. Future semiconductor devices in integrated circuits will soon be developed into a size as small as 45 nm generation, and maybe the porous SiOCH base film4 will be applied in semiconductor industry. The basic mechanical properties of the Young’s Modulus and hardness which are for semiconductor industry are about 12 GPa and 1.5 Gpa, respectively. The ultra low
dielectric constant materials must be lower than 2.0 for the next ULSI generation. Both porous silica (Xerogels/Aerogels) and PTFE polymer (Teflon) materials conform to the ultra low k requirement. The porous silica has high hardness and hydrophilic characteristic. These properties are disadvantageous for IC process integration of low k materials. The PTFE polymer has hydrophobic characteristic. PTFE films can be deposited by sputtering, but the sputtering yield is very low. The mixing of fluorocarbon and hydrocarbon gases produces fluorinated amorphous carbon (a-C:F) film by CVD methods. The a-C:F film properties are similar to PTFE. It accelerates the deposition rate and decreases the dielectric constant as well.
With its lower dipole moment3 and lower electronic polarizability5 properties, the C-F bonds have a large electronegativity, therefore considered ideal to construct low dielectric material. Savage et al. were the first one to deposit a-C:F film by plasma CVD in 19756, and Endo et al.
were the first one who suggested fluorinated amorphous carbon thin film as a low dielectric constant material in 19957. The permanent dipole moment of fluorinated carbon molecule is as low as 0.53, while the electronic polarizability of fluorocarbon bond is as low as 0.565. Thus, the a-C:F film is an excellent low dielectric constant material.
Endo et al. used helicon plasma enhanced CVD to deposit a-C:F film8. This particular CVD brings up the deposition rate (0.15 ~ 0.3 µm/min), and accelerates the structural cross-linking in the film. Such network structure improves the thermal stability. All of the sp3 bounds of PTFE polymer are linear CF2 bonds, and the thermal stability temperature of PTFE polymer is as high as 200℃. However, some of network structure
is in the a-C:F film, and its thermal stability temperature can be over 300℃.
Unfortunately the highly network structure of a-C:F film will cause the dielectric constant to grow. Endo et al.9 and Yokomichi et al.10 found that nitrogen atom can help to form the cross-linking in the a-C:F:N film, of which dielectric constant is ~ 2.4 until 400℃. If we increases the fluorinate concentration in a-C:F films, we can get ultra low dielectric constant (~1.5) in a-C:F films. And then we can use hydrogen or nitrogen plasma treatment to enhance the thermal stability. In addition, a-C:F film is an ultra low k film, the a-C:F film is also an excellent field emission material11. The work function of a-C:F film is ~0.012 eV. The a-C:F film has a lot of advantages.
For example, the a-C:F film has high break-down voltage, hydrophobic property, and is easy for etching.
Table 2-1. Low dielectric constant materials for ULSI interconnection.
Table 2-2. Basic requirements for low dielectric constant materials.
2-2 Fluorinated Amorphous Carbon
Fluorocarbon film was used as a surface coating in early stage because fluorocarbon film has lower frication and hydrophobic characteristics12. Even though the fluorocarbon films are always produced after chamber cleaning or silicon dioxide etching in semiconductor industry13-16, but the fluorocarbon did not serve as low dielectric materials in integrated circuits, since the dielectric constant of silicon dioxide material was low enough for very large scale integrated (VLSI) device at that moment. However, when the semiconductor was promoted to the 180 nm ultra large scale integrate (ULSI) devices, the dielectric constant of low dielectric constant (low k) material must be lower than 3.017.
Therefore, there was a demand for new interlayer insulators to improve the switching performance of ultra-large-scale integrated (ULSI) devices. The need has led to intensive studies on several low-dielectric materials. Among them, fluorinated amorphous-carbon (a-C:F) films have received most attention due to their thermal stability and low dielectric constant7, 18-24. The relationship between these properties and the film compositions has been discussed in several studies21-25. The stability and dielectric properties could be manipulated by adjusting the F content of the films via control of the flow rate ratio of the source gases or by other techniques. A higher F concentration in the films normally reduces the thermal stability and the dielectric constant because it leads to weaker C-C cross linking, lower density and reduced polarization26-28. However, the films with high excess F, despite their thermal instability, have low leakage current and extremely high breakdown resistance. In a low
temperature environment, they could be applied as excellent insulating layers. In this thesis, we report on the structural and electrical properties of the high-F-content a-C:F films prepared by electron cyclotron resonance/chemical-vapor deposition (ECR-CVD).
In fluorocarbon plasma, it is well known that low F/C ratios of the precursor gases usually facilitate film growth, while high ratios induce erosion processes29. Previous studies on plasma CVD using CH4 and CF4 gases18-24 showed that the deposition rate of a-C:F films first increases with the ratio of F/C, then declines rapidly from the maxima to zero, where the ratio was calculated in consideration of the total number of F atoms divided by the total number of C atoms in the inlet gas flux per unit time. Apparently, the growth rate of a-C:F films of high F content using the source gases of very high F/C ratio was limited by the etching effect of concentrated F radicals. On the other hand, a much higher deposition rate of a-C:H films by PECVD can be achieved by using C2H2 as a precursor rather than CH430. Therefore, CF4 and C2H2 were employed as the source gases in this study to ensure a reasonable film growth rate.
The microwave plasma ECR-CVD system described in this thesis was used to produce reactive chemical species, and the RF bias assisted the precursor to impinge on the substrate, resulting in the deposition of the film.
Both C2H2 and CF4 were the precursors used for the synthesis of the a-C:F films. The C2H2 gas has a high C/H ratio, and could contribute enough carbon source to support a film network to increase the deposition rate31. The CF4 gas has a high F/C ratio enough to supply sufficient fluorocarbons to raise the concentration of fluorine in the a-C:F film. The CF4 plasma contains mainly CF3+ ions, F neutrals, CFn radicals and negative ion species.
Fluorine atoms control not only the concentration of CFn radicals through the gas-phase reactions but also the surface reactions32-36.
In this thesis, we report on the chemical and physical properties of the fluorinated amorphous carbon film prepared by ECR-CVD, at CF4 flow ratios from R = 0.90 to R = 0.98, where R is CF4/[CF4+C2H2]. This is to compare the difference between the as-deposited film and the film after annealing at 300oC for 30 min. For this purpose we used the analytical instruments as follows: Fourier transform infrared absorption spectrophotometer (FTIR), electron spectroscopy for chemical analyzer (ESCA), high resolution electron energy loss spectrometer (HREELS), electron spin resonance spectrometer (EPR), photoluminescence spectrometer (PL), ultraviolet / visible spectrophotometer (UV/VI), field emission scanning electron microscope (FESEM), pulse laser spectrometer and atomic force microscope (AFM). Current-voltage (I-V) measurements were performed with the HP4156, and the capacitance (C-V) characteristic was measured at a frequency of 1MHz using the HP4280. The sp2 content and the fluorine concentration would affect the photoluminescence lifetime, as well as energy band gap of the a-C:F films. With the rise of the temperature, the dangling bond density increases, resulting from the growing numbers of unpaired spins in the defects in the films.
Reference
1 C. R. Savage, R. B. Timmons and J. W. Lin, Adv. Chem. Ser., 236, 745 (1993).
2 Kazuhiko Endo, Toru Tatsumi, J. Appl. Phys., 78, 1370 (1995).
3 CRC Handbook of Chemistry and Physics. 77th Ed., CRC Press, Boca Raton (1996).
4 A. Grill and V. Patel, J. ElCh. Soc., 151, F133 (2004).
5 K. J. Miller, J. Am. Chem. Soc., 112, 8533 (1990).
6 K. G. Budinski, J. Vac. Sic. Technol. 12, 786 (1975).
7 K. Endo, T. Tatsumi, J. Appl. Phys., 78, 1370 (1995).
8 K. Endo and T. Tatsumi, Appl. Phys. Lett., 68, 2864 (1996).
9 K. Endo and T. Tatsumi, Appl. Phys. Lett., 68, 3656 (1996).
10 H. Yokomichi, and A. Masuda, J. Non-Cryst. Solids 271, 147 (2000) 641.
11 S. H. Lai, K. P. Huang, Y. M. Pan, Y. L. Chen, L. H. Chan, P. Lin, and H. C. Shih, Chem. Phys. Let. 382, 567 (2003).
12 C. R. Savage, R. B. Timmons and J. W. Lin, Adv. Chem. Ser., 236, 745 (1993).
13 R. d’Agostino, R. Lamendola, P. Favia, and A. Giqel, J. Vac. Sci.
Technol. A 12, 308 (1994).
14 G. S. Oehrlein, Y. Zhang, D. Vender, and M. Haveriag, J. Vac. Sci.
Technol. A 12, 323 (1994).
15 M. J. Sowa, M. E. Littau, V. Pohray, and J. L. Cecchi, J. Vac. Sci.
Technol. A 18, 2112 (2000).
16 E. A. Joseph, B. Zhou, S. P. Sant, L. J. Overzet, and M. J. Goeckner, J.
Vac. Sci. Technol. A 22, 689 (2004).
17 “The National Technology Roadmap for Semiconductors”, Semiconductor Industry Association, San Jose, CA, 1997.
18 H. Yokomichi, T. Hayashi, T. Amano and A. Masuda: J. Non-Cryst.
Solids 227, 641 (1998).
19 Y. Ma, H. Yang, J. Guo, C. Sathe, A. Agui and J. Nordgren: Appl.
Phys. Lett. 72, 3353 (1998).
20 H. Yang, D. J. Tweet, Y. Ma and T. Nguyen: Appl. Phys. Lett. 73, 1541 (1998).
21 J. A. Theil: J. Vac. Sci. & Technol. B 17, 2397 (1999).
22 K. Endo and T. Tatsumi: Jpn. J. Appl. Phys. 36, L1531 (1997).
23 H. Yokomichi and A. Masuda: J. Appl. Phys. 86, 2468 (1999).
24 L. G. Jacobsohn, D. F. Franceschini, M. E. H. Maia da Costa and F. L.
Freire, Jr. : J. Vac. Sci. & Technol. A 18, 2230 (2000).
25 T. W. Mountsier and J. A. Samuels: Thin Solid Films 332, 362 (1998).
26 H. Yokomichi, T. Hayashi and A. Masuda: Appl. Phys. Lett. 72, 2704 (1998).
27 K. Endo and T. Tatsumi: J. Appl. Phys. 86, 2739 (1999).
28 J. P. Chang, H. W. Krautter, W. Zhu, R. L. Opila and C. S. Pai: J. Vac.
Sci. & Technol. A. 17, 2969 (1999).
29 J.W. Coburn and H. F. Winters: J. Vac. Sci. & Technol. 16, 391 (1979).
30 P.S. Andry, P.W. Pastel and W. J. Varhue: J. Mater. Res. 11, No.1
(1996) 221.
31P. S. Andry, P. W. Pastel, and W. J Varhue, J. Mater. Res. 11, 221 (1996).
32 G. Chuge and J. P. Booth, J. Appl. Phys. 85, 3952 (1999).
33 W. Schwarzenbach, G. Cunge, and J. P. Booth, J. Appl. Phys. 85, 7562 (1999).
34 K. Teii, M. Hori, M. Ito, T. Goto, and N. Ishii, J. Vac. Sci. Technol. A 18, 1 (2000)
35 L. G. Jacobsohn, D. F. Franceschini, M. E. H. Maia da Costa, and F. L.
Freire, Jr, J. Vac. Sci. Technol. A 18, 2230 (2000).
36 K. Sasaki, H. Furukawa, K. Kadota, and C. Suzuki, J. Appl. Phys. 88, 5585 (2000).