Magnetized Inductively Coupled Plasma Research Articles

Dry Etching Characteristics of 16-nm Amorphous Carbon Layer in a Dual-Frequency Plasma Etcher

The etching characteristics of a patterned amorphous carbon layer (ACL) in a magnetized inductively coupled plasma (MICP) etcher were investigated. The etch rate and etched profile dependences on the source power, the neutral gas pressure and the dual bias power ratio were investigated. We measured the ion energy distribution (IED), the ion flux, plasma density and electron temperature using an ion energy analyzer (IEA) and a dual Langmuir probe (DLP). The etching rate and the anisotropy of ACL relative to change in the dual bias power ratio of two bias power generators operating at frequencies of 2 and 13.56 MHz were observed. The bowed profile could be improved by appropriate adjustment of the dual bias power ratio.

A Study on Uniformity Characteristics of a Magnetized Inductively Coupled Plasma

Non-uniformities in the plasma parameters in a magnetized inductively coupled plasma (M-ICP) etcher were investigated. Further in text of this article the silicon dioxide and the photoresist are called as abbreviations the oxide and the PR accordingly. The etch results of an oxide and PR were correlated with the non-uniformity characteristics of the plasma parameters. The plasma density non-uniformities and the oxide etch-rate non-uniformities in M-ICP-V (9.28 and 2.4%, respectively) were lower than those in M-ICP-A (14.6% and 21.4%, respectively) or ICP (13.03 and 5.2%, respectively). The profile angle of the etched silicon dioxide in ICP, M-ICP-A, and M-ICP-V configurations were also measured and the angle in M-ICP-V was approximately 85° and similar in ICP.

Study on spatial distribution of plasma parameters in a magnetized inductively coupled plasma

Spatial distributions of various plasma parameters such as plasma density, electron temperature, and radical density in an inductively coupled plasma (ICP) and a magnetized inductively coupled plasma (M-ICP) were investigated and compared. Electron temperature in between the rf window and the substrate holder of M-ICP was higher than that of ICP, whereas the one just above the substrate holder of M-ICP was similar to that of ICP when a weak (<8 G) magnetic field was employed. As a result, radical densities in M-ICP were higher than those in ICP and the etch rate of oxide in M-ICP was faster than that in ICP without severe electron charging in 90 nm high aspect ratio contact hole etch.

Low-k material damage during photoresist ashing process

The change of –OH and –CH3 component ratios in Fourier transform-infrared analysis of low-k materials during photoresist (PR) ashing processes were compared to assess the differences in the damages to low-k materials in a reactive ion etch (RIE) chamber and a magnetized-inductively coupled plasma (M-ICP) chamber. In M-ICP, the PR ashing rate was 28.1% higher than that of RIE, but the low-k material damage in M-ICP decreased when typical ashing conditions were used in each machine. The dependences of low-k material damage and PR ashing rate on the pressure, source power, and bias power in the M-ICP chamber were studied. We measured the ion energy distributions using an ion energy analyzer from which the flux could be also obtained. We found that the PR ashing rate increased as the ion flux increased, while the low-k material damage also increased as the ion flux and the incident ion energy increased. However, as the pressure decreased, the ion flux increased dramatically and the ion energy decreased. As a result, the PR ashing rate could be high and the low-k material damage low.

A study on reactive ion etching lag of a high aspect ratio contact hole in a magnetized inductively coupled plasma

Reactive ion etching lag (RIE lag) characteristics of a high aspect ratio contact hole in a magnetized inductively coupled plasma (M-ICP) etcher were investigated. The dependence of RIE lag on various process parameters, such as neutral gas pressure, magnetic flux density, bias power, bias frequency, and source power were investigated. It was confirmed that RIE lag could be reduced by properly adjusting these variables. Furthermore, the application of a magnetic field to the ICP etcher was helpful to increase the oxide etch rate and oxide-to-amorphous carbon layer selectivity.

Reduced electron temperature in a magnetized inductively-coupled plasma with internal coil

The effect of magnetic filtering on the electron energy distribution function is studied in an inductive discharge with internal coil coupling. The coil is placed inside the plasma and driven by a low-frequency power supply (5.8 MHz) which leads to a very high power transfer efficiency. A permanent dipole magnet may be placed inside the internal coil to produce a static magnetic field around 100 Gauss. The coil and the matching system are designed to minimize the capacitive coupling to the plasma. Capacitive coupling is quantified by measuring the radiofrequency (rf) plasma potential with a capacitive probe. Without the permanent magnet, the rf plasma potential is significantly smaller than the electron temperature. When the magnet is present, the rf plasma potential increases. The electron energy distribution function is measured as a function of space with and without the permanent magnet. When the magnet is present, electrons are cooled down to low temperature in the downstream region. This region of low electron temperature may be useful for plasma processing applications, as well as for efficient negative ion production.

Effect of bias application to plasma density in weakly magnetized inductively coupled plasma

Independent control of the ion flux and energy can be achieved in a dual frequency inductively coupled plasma (ICP) system. Typically, the plasma density is controlled by the high-frequency antenna radio-frequency (RF) power and the ion energy is controlled by the low-frequency bias RF power. Increasing the bias power has been known to cause a decrease in the plasma density in capacitively coupled discharge systems as well as in ICP systems. However, an applied axial magnetic field was found to sustain or increase the plasma density as bias power is increased. Measurements show higher electron temperatures but lower plasma densities are obtained in ordinary ICP systems than in magnetized ICP systems under the same neutral gas pressure and RF power levels. Explanations for the difference in the behavior of plasma density with increasing bias power are given in terms of the difference in the heating mechanism in ordinary unmagnetized and magnetized ICP systems.

Self-consistent simulation study on magnetized inductively coupled plasma for 450 mm semiconductor wafer processing

The characteristics of weakly magnetized inductively coupled plasma (MICP) are investigated using a self-consistent simulation based on the drift–diffusion approximation with anisotropic transport coefficients. MICP is a plasma source utilizing the cavity mode of the low-frequency branch of the right-hand circularly polarized wave. The model system is 700mm in diameter and has a 250mm gap between the radio-frequency window and wafer holder. The model chamber size is chosen to verify the applicability of this type of plasma source to the 450mm wafer process. The effects of electron density distribution and external axial magnetic field on the propagation properties of the plasma wave, including the wavelength modulation and refraction toward the high-density region, are demonstrated. The restricted electron transport and thermal conductivity in the radial direction due to the magnetic field result in small temperature gradient along the field lines and off-axis peak density profile. The calculated impedance seen from the antenna terminal shows that MICP has a resistance component that is two to threefold higher than that of ICP. This property is practically important for large-size, low-pressure plasma sources because high resistance corresponds to high power-transfer efficiency and stable impedance matching characteristics. For the 0.665Pa argon plasma, MICP shows a radial density uniformity of 6% within 450mm diameter, which is much better than that of nonmagnetized ICP.

Effective viscosity model for electron heating in warm magnetized inductively coupled plasma discharges

An effective viscosity model for warm magnetized inductively coupled plasma (MICP) discharges has been derived. It calculates the power absorbed inside MICP discharges that takes nonlocal behavior into account with the help of effective viscosity terms in the momentum equations for right-handed and left-handed components of the wave. The validity of this model for warm MICP discharges has been checked by comparing it with self-consistent kinetic model for warm MICP discharges. This effective viscosity model shows nonmonotonic decay for right- and left-handed components of the electric field inside MICP discharges. It also shows regions of negative power absorption which cannot be shown using conventional fluid models. The power absorbed per unit area (for right- and left-handed components) calculated using effective viscosity model is similar to that calculated using computationally extensive kinetic models over a wide range of MICP discharge conditions.

Effect of electron thermal motion on plasma heating in a magnetized inductively coupled plasma

Power absorbed inside the magnetized inductively coupled plasma (MICP) is calculated using three different warm MICP models and is then compared with the result of the cold MICP model. The comparison shows that in the propagating region (ω<∣Ωe∣), under the cavity resonance conditions, warm plasma heating Swarm is significantly less than the cold plasma heating Scold, unless the distance traveled by the electrons due to their thermal motion, during the effective wave period, becomes significantly less than the wavelength of the cavity wave. Furthermore, in the propagating region, when ω≈∣Ωe∣, there appears a valley on the plot of η(ω)=Swarm∕Scold versus ω showing the negative effect of electron thermal motion on plasma heating. This valley widens and gets smoother with an increase in the plasma length. In the nonpropagating region (ω>∣Ωe∣), the maximum value of η(ω) exists when ω−∣Ωe∣≈vth∕δ, showing that, in the presence of the external magnetic field, the thermal motion of the electrons leads to a Doppler shift of the frequencies, at which collisionless heating is the dominant mode of electron heating. Furthermore, in the nonpropagating region, when ω≈∣Ωe∣, the skin depth of the right circularly polarized electric field decreases with magnetic field. This decrease in the skin depth results in an increase of collisionless heating under the Doppler-shifted wave particle resonant condition of ω−∣Ωe∣≈vth∕δ. It is also observed that, for large plasma length, the results of all the three warm MICP models are consistent with each other.

Heating mechanism in one-dimensional bounded magnetized inductively coupled plasma

A self-consistent nonlocal kinetic model has been established to investigate the electron heating mechanism in an one-dimensional bounded magnetized inductively coupled plasma (MICP) under low pressures. The interaction function of electrons with rf electric field and electron energy distribution function (EEDF) are determined by solving the linearized Boltzmann equation incorporating with the Maxwell equations. The numerical results show that the presence of the direct-current (dc) magnetic field plays an important role in the EEDF and high-energy electrons are efficiently heated by the Azbel–Kaner resonances under the anomalous skin-effect conditions.

Wave stimulated phenomena in inductively coupled magnetized plasmas

Linear and nonlinear wave phenomena and their influence on the discharge performance are considered in the helicon plasma (HP) and magnetized inductively coupled plasma (MICP). Magnetic field configuration is shown to result in strong variation in the efficiency of plasma production caused by alterations of the character of wave processes rather than by the particle confinement. Effects of magnetic configuration are found to be dominant in operation of the compact HP with permanent magnet. Low frequency turbulence in the megahertz range is revealed to be inherent for both the HP and MICP and to depend strongly on the magnetic configuration. Enhanced axial electric field arising in front of the metal end plate in the HP is argued to be a potential source of the nonlinear effects.

Radio frequency potential of inductive plasma immersed in a weak magnetic field

Radio frequency potential has been revealed in inductively coupled plasma (ICP) immersed in a weak magnetic field. The rf plasma potential measured in a wide range of driving frequency and argon pressure reached tens of volts at a relatively weak magnetic field and was found falling with an increase in the driving frequency and gas pressure. The appearance of rf plasma potential in magnetized ICP is interpreted as a result of rf Hall effect caused by the electron rf drift in a crossed magnetic field.

Experimental measurement of the electron energy distribution function in the radio frequency electron cyclotron resonance inductive discharge.

Recently, the existence of electron cyclotron resonance (ECR) in a weakly magnetized inductively coupled plasma (MICP) has been evidenced [ChinWook Chung et al., Phys. Rev. Lett. 80, 095002 (2002)]. The distinctive feature of the ECR effect in the MICP is efficacious heating of low-energy electrons. In the present paper, electron heating characteristics in the MICP have been investigated by observing electron energy distribution function dependencies on various external parameters such as gas pressure, driving frequency, and rf power (electron density). It is found that the ECR effect on electron heating becomes enhanced with decreasing pressure or increasing driving frequency. The ECR heating becomes weak at high rf power due to the electron-electron collisions.

Etch characteristics of SrBi 2Ta 2O 9 (SBT) thin films using magnetized inductively coupled plasmas

In this study, SrBi 2Ta 2O 9 (SBT) thin films were etched using a magnetized inductively coupled plasma (MICP) and their etch characteristics were investigated as a function of Cl 2/Ar gas mixing ratio, inductive power and bias voltage. The SBT etch rate appeared to show the highest value when Cl 2 concentration in the Cl 2/Ar mixture was 30%, even though the SBT etch rate remained nearly constant when the Cl 2 concentration in the mixture was higher than 10%. SBT etch rates were also strongly influenced by both inductive power and bias voltage. This result implies that the etching of SBT is dependent upon both ion bombardment and chemical reaction. An etch rate of 150 nm/min could be obtained using 30% Cl 2/70% Ar, 6.7 Pa of operation pressure, 600 W of inductive power, and −300 V of bias voltage. The etch selectivity of SBT over Pt, and photoresist were less than 0.5 and 1.0∼1.2, respectively, when the Cl 2 was more than 10% in Cl 2/Ar mixtures. An SBT (200 nm)/Pt (100 nm)/Ti (100 nm) heterostructure was etched using the MICP with 30% Cl 2/70% Ar mixture, which resulted in an anisotropic etch profile without the formation of sidewall residue.

Inductively coupled plasma heating in a weakly magnetized plasma

A one-dimensional analysis of electron heating process in a weakly magnetized, inductively coupled plasma (MICP) is presented. It is found that the main difference in the heating process of a MICP from that of a usual unmagnetized ICP is in that circularly polarized wave modes can exist in the plasma. The right handed circularly polarized wave (R-wave) can propagate into the plasma and its amplitude can be enhanced by cavity resonance effect at an appropriate chamber length and external magnetic field strength. The enhanced R-wave amplitude can raise the heating efficiency significantly. It is also found that a bounce cyclotron-resonance effect can exist, which, however, is not as significant as the cavity resonance effect.