On phenomena in ionized gases
Simulation of Plasma Processing with FPS3D
Download 9.74 Mb. Pdf ko'rish
|
- Bu sahifa navigatsiya:
- 1. Introduction - FPS3D.
- 2. Two Examples of FPS3D Simulations
- 2.1. Etching/Implantation with Ar/Cl 2 plasma
- 2.2. SiN Atomic Layer Deposition with Cycles of Dichlorosilane and Ammonia Plasma
- References
- Determination of collisional quenching rate coefficients of metastable excited atoms Ar( 3 P 2
- 2. Experimental apparatus and method
- 4. References
- Rate equation analysis of ROS/RNS in plasma-treated water
- 2. Calculation methods and conditions
- 3. Results and discussion
- A modified fluid simulation of an inductively coupled plasma discharge with radio frequency bias considering heat transfer effect
Simulation of Plasma Processing with FPS3D
P. Moroz P 1 P , U D. J. Moroz UP 2 P
P 1 P
P
P School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA
Simulation of plasma processing via a Monte Carlo feature-scale simulator FPS3D is demonstrated by two very different examples: simulation of etching and implantation for the case of Ar/Cl 2
plasma, and simulation of atomic layer deposition with a cyclic process of dichlorosilane gas followed by the ammonia plasma. Comparison with experiments is provided.
Feature-scale simulations allow reasonably fast simulation of etching or deposition profiles. These simulations are many orders of magnitude faster that the higher accuracy MD methods. There have been many types of feature-scale simulators developed since the 1970s, and their capabilities differ significantly from each other. FPS3D is a general software package designed to be applicable to any situation met by the semiconductor industry, be that etching, deposition, implantation, atomic layer processing, etc., or any combination of these. FPS3D uses a cellular model for representing solid materials, but that model goes well beyond the traditional approach. In FPS3D, each cell can contain different molecules and the number of molecules per cell is not fixed but rather is determined by the volume of the cell and by the size of molecules it contains. Also, FPS3D uses Monte Carlo pseudo-particles for representing all incoming fluxes. These particles are launched such that statistically they represent given angle-energy distributions of all relevant fluxes of species coming to the surface. Each particle typically contains many molecules, but preferably significantly fewer than the number of molecules in a full cell. Upon collision of such a particle with a solid material, the code determines a set of involved cells where interactions occur and computes the output on the basis of user-specified reactions. FPS3D is free from many limiting assumptions of most other feature-scale simulators. For etching it allows simultaneously calculation or implantation, and for deposition, it allows materials to grow in accordance with their density, and in accordance with the actual sizes of participating molecules.
FPS3D can be applied to features of different scales ranging from nanometers 3 to microns 2 . By selecting proper sizes of material cells and incoming particles, the code obtains reasonable accuracy within reasonable calculation time.
To setup surface reactions for the case of Ar/Cl 2
ion energies of 35, 55 and 75 eV. Results of simulations were able to reproduce the experimental data 4
energy ions. We have used the same chemistry for the case with high energy Ar+ ions of 1.5 keV. Energetic ions penetrate deep into material, and sometimes could lead to plasma-induced damage. We simulated this process both in 2D and 3D.
Dichlorosilane and Ammonia Plasma We considered a case of SiN ALD that uses cycles of dichlorosilane gas deposition followed by the nitration by ammonia plasma. This case is interesting because it results in the deposition of 1 ML only after two cycles. We believe that one of the main reasons for this is steric hindrance by SiH 2 Cl 2
molecules. FPS3D is designed to take steric hindrance into account. This large molecule could cover two interaction sites on the surface, preventing those sites from interacting with other molecules until the large molecule reacts and reaction sites become accessible again. Results of simulations are compared with experiments 5 .
References [1] P. Moroz, D.J. Moroz, ECS Transactions, 50 61 (2013). [2] P. Moroz, D. J. Moroz, J. Physics: CS 550 (2014) 012030 (2014). [3] P. Moroz, D. J. Moroz, to be published in Japan. J. Appl. Phys. (2017). [4]
J. P. Chang, A. P. Mahorowala, H. H. Sawin, J. Vac. Sci. Technol. A 16, 217 (1998).
[5] H. Goto, K. Shibahara, S. Yokoyama, Appl. Phys. Lett. 68, 3257 (1996). Topic 14 157
XXXIII ICPIG, July 9-14, 2017, Estoril/Lisbon, Portugal
Determination of collisional quenching rate coefficients of metastable excited atoms Ar( 3 P 2 ) by Ar and H 2 O
U S. Suzuki UP 1 P , Y. Usui 1 , H. Itoh P 1 P (Chiba Institute of Technology)
P 1 P
Narashino, Chiba 275-0016, Japan
We observed the transient current after turning off the UV light illuminating a cathode used to supply the photoemission current in a Townsend discharge to measure the effective lifetime of the metastable excited atoms Ar( 3 P 2 ). The diffusion coefficient of Ar( 3 P
) in argon, the reflection coefficient of Ar( 3 P 2 ) at the electrode surface and the collisional quenching rate coefficient of Ar( 3 P 2 ) by ground-state atoms Ar( 1 S
) were determined from the observed effective lifetime of Ar( 3 P
). Moreover, the collisional quenching rate coefficient of Ar( 3 P
) by H 2 O was also determined.
In this study, the three fundamental coefficients of metastable excited atoms Ar( 3 P
) were determined by a non-spectroscopic measurement and curve fitting to theoretical values of the effective lifetime derived from the solution of the diffusion equation. 2. Experimental apparatus and method Details of the experimental apparatus and the procedure used for numerical analysis have already been reported [1,2]. The purity of Ar gas used in the experiment was 99.999%.
In figure 1, the four solid lines show the observed effective lifetime τ 1 of Ar( 3 P 2 ) in Ar plotted against the gas pressure for different gap lengths. From the results, we determined the diffusion coefficient D m1
3 P 2 ), the collisional quenching rate coefficient k of Ar( 3 P 2 )
by Ar( 1 S 0 ) and the reflection coefficient R of Ar( 3 P 2 ) at the electrode to be 42.2±1.4 cm 2 /s,
(2.96±0.05)×10 -15
cm 3 /s and 0.13±0.02, respectively. The value of k is consistent with that reported by Molnar [3] and also with values previously obtained by spectroscopic measurement [4-7]. The
temperature dependence of k(T) was also derived as k=3.9×10 -7 exp(-5700/T) [300≦T≦343 K]. On the basis of the results, the experiments were extended to an Ar/H 2 O(112 ppm) mixture and the preliminary results are shown in Fig. 1. The collisional quenching rate coefficient k’ of Ar( 3 P
) by H 2 O was determined to be 2.3×10 -10
cm 3 /s, which is 10 5 times larger than the value of k. Our value k’ was consistent with those reported by Bourene and Le Calvè [8] and Balamuta and coworkers [9, 10] of 1.84×10 -10
cm 3 /s and 2.16×10 -10 cm
3 /s, respectively, who described the generation of O, H and OH as the by-products of H 2 O.
4. References [1] S.Suzuki and H.Itoh, J. Phys. D: Appl. Phys. 49 (2016) 185202. [2] S.Suzuki, H.Itoh, N.Ikuta and H.Sekizawa, J. Phys. D: Appl. Phys. 25 (1992) 1568-1573. [3] J. P. Molnar, Phys. Rev. 83 (1951) 940-952. [4] A. H. Furch and F. A. Grant, Phys. Rev. 104 (1956) 356-361. [5] E. Ellis and N. D. Twiddy, J. Phys. B: At. Mol. Phys. 2 (1969) 1366-1377. [6] A.V. Phelps and J. P. Molnar, Phys. Rev. 89 (1953) 1202-1208. [7] R. A. Gutcheck and E. C. Zipf Bull, Am. Phys. Soc. 17 (1972) 395. [8] M. Bourène and J. Le Calvé, J. Chem. Phys. 58 (1973) 1452. [9] J. Balamuta and M.F. Golde, J. Chem. Phys. 76 (1982) 2430. [10] J. Balamuta, M.F. Golde and Yueh-Se Ho, J. Chem. Phys. 79 (1983) 2822. Topic number 1 Gas pressure p 0 (Torr)
E ff ec ti v e li fe ti m e t 1
(m s) d=1.5cm d=1.2cm d=1.0cm d=0.7cm Ar Ar/H 2 O(112ppm) 0.1 1
100 0.1
1 10
Fig. 1 Effective lifetime of metastable excited atoms Ar(
3 P 2 ) in Ar. 158
XXXIII ICPIG, July 9-14, 2017, Estoril/Lisbon, Portugal
Rate equation analysis of ROS/RNS in plasma-treated water
K. Takahashi P 1 P , S. Kawaguchi UP 1 P , K. Satoh 1 , H. Kawaguchi 1 , I. Timoshkin 2 , M. Given 2 , and S. MacGregor 2
P 1 P
P
P
2 O 2 , NO 2 - , and NO 3 - as reactive oxygen species and reactive nitrogen species in water, exposed to a pulsed discharge in nitrogen atmosphere, are calculated using rate equations coupled with acid-base equilibrium between NO 2 - and HNO 2 and a chemical reaction between H 2 O 2 and HNO
2 , considering the temporal variations of flux rates of H 2 O
, NO 2 - , and NO 3 - , which is estimated from measurement data in our previous work. Furthermore, the calculated concentrations are fitted to the measurement values. It is found that the calculated concentrations are in approximate agreement with the measured data. It is also found that the generation rates of H 2 O
, NO 2 - and NO 3 - are estimated to be 9.5×10 -7 , 4.5×10
-7 , and 2.5×10 -7 M/s.
1. Introduction In recent years, plasma-treated water, in which reactive oxygen species and reactive nitrogen species (ROS/RNS) dissolve, has gained increasing attention, because ROS/RNS in the plasma-treated water, such as H
2 O 2 (hydrogen peroxide), NO 2 - (nitrite ion), HOONO (peroxynitrous acid), and NO 3 -
and/or synergistic effects between these species play a key role in various applications such as disinfection [1] and plant growth promoting [2]. To selectively and/or effectively produce the ROS/RNS and utilize the plasma-treated water effectively and efficiently, it is important to clarify the generation process of the ROS/RNS; however, the generation process has not yet been clarified. In this work, we calculated the ROS/RNS concentrations in water, exposed to a pulsed discharge in nitrogen atmosphere, by solving rate equations coupled with acid-base equilibrium, and fitted the results to measurement values in our previous work [3].
H 2
2 , NO
2 - and NO 3 - are produced in the water, and the pH of the water decreases with plasma exposure. NO 2 -
HNO 2 (nitrous acid), and HNO 2 reacts with H 2 O
to form NO
3 -
as shown in Eq. (1) [4]; therefore, the concentrations of H 2 O
, NO 2 - , NO 3 - , and HNO 2 as functions of time t are expressed by rate equations as shown in Eqs. (2)-(7). H 2
2 + HNO
2 + H
+ → NO
3 - + H 2 O + 2H
+
(1) (2)
(3)
(4)
(5)
(6)
(7)
Here, [M], F(t), G M , k, and f M represent the concentration of product M, the time variation of flux, the generation rate of product M, rate constant, and the abundance ratio of product M, respectively. The time variation of flux is estimated from the differential coefficient of the measured concentration variations of NO 2 -
3 - , and the abundance ratio is calculated from the pH of water, which is the measured data. The concentrations of H 2 O
, NO 2 - , and NO
3 - are calculated using the 4th order Runge- Kutta method, and fitted to the measurement values by varying G M and k. 3. Results and discussion The calculated concentrations of H 2 O
, NO 2 - and NO 3 - as functions of time are in approximate agreement with the measurement values. Therefore, NO 2 - is converted into HNO 2 under acidic conditions, and then HNO 2 reacts with H 2 O 2 to form NO 3 - . It is found that the generation rates of H 2 O
, NO 2 - and NO 3 - for the plasma treated water are estimated to be 9.5×10 -7
-7 , and 2.5×10 -7 M/s.
4. References [1] A. Kojtari, T.K. Ercan, J. Smith, G. Friedman, R.B. Sensenig, S. Tyagi, S.G. Joshi, H-F. Ji, and A.D. Brooks, J. Nanomed. Biother. Discov. 4 (2013) 120. [2] K. Takaki, J. HTSJ, 51 (2012) 64. [3] K. Takahashi, K. Satoh, H. Itoh, H. Kawaguchi, I. Timoshkin, M. Given, and S. MacGregor, Jpn. J. Appl. Phys. 55 (2016) 07LF01. [4] G. Merényi, J. Lind, G. Czapski, and S. Goldstein, Inorg. Chem. 42 (2003) 3796. Topic Number: 17 ] H ][ HNO
][ O H [ ) ( d ] O d[H 2 2 2 O H 2 2 2 2 k G t F t ] H ][ HNO
][ O H [ ) ( d ] d[NO 2 2 2 NO 3 3
G t F t ] H ][ HNO
][ O H [ d ] d[HNO 2 2 2 2 k t ]) HNO [ ] NO ([ ] [NO 2 2 NO 2 2 f ]) HNO [ ] NO ([ ] [HNO 2 2 HNO 2 2
2 NO 2 ) ( d ] d[NO
G t F t 159
XXXIII ICPIG, July 9-14, 2017, Estoril/Lisbon, Portugal
A modified fluid simulation of an inductively coupled plasma discharge with radio frequency bias considering heat transfer effect
Y. D. Jeong P 1 P , U Y. J. Lee UP 1 P , D. C. Kwon P 2 P , H. H. Choe P 1
P 1 P
P
P
The plasma characteristics in an inductively coupled plasma (ICP) discharge with radio frequency bias (RF) were investigated. A two-dimensional axisymmetric structure was simulated by using a modified fluid model. Large and multi-size zones were used for the calculations of the Two-Term Boltzmann approximation electron energy distribution function (EEDF) and ion energy distribution function (IEDF) calculated by using spatial averaged plasma parameters. The energy and mobility of ion were calculated by using the IEDF at each zone. In addition, the heat transfer was considered. Voltage drop across the coils due to the reactance were considered that the capacitive field effect of the antenna was also considered. Effects of these application were analysed.
1. Introduction Simulations for an inductively coupled plasma (ICP) Argon discharge with radio frequency (RF) bias for semiconductor device processes were conducted and
plasma characteristics were investigated. Although the fluid model, one of the models describing the plasma, has some accuracy problems, it provides relatively rapid computation and satisfactory solution for
moderate pressure conditions [1-2]. The simulation model is based on this.
In this study, a two-dimensional axisymmetric structure was used for the simulation. Voltage drop of the antenna coil is considered and this effect was investigated. In addition, the electron energy distribution function (EEDF) was calculated. The EEDF is a dominant factor for determining plasma characteristics, since it is a key parameter in calculations of electron transport properties (e.g., the electron mobility, the electron diffusivity, reaction coefficients). Ion temperature and mobility were computed using particle tracing mechanism. The effects of these applications were studied.
Download 9.74 Mb. Do'stlaringiz bilan baham: |
ma'muriyatiga murojaat qiling