Helicon Source Research Articles

Volume and surface rf power absorption in a helicon plasma source

We consider the excitation and absorption of waves in a plasma cavity with parameters typical for helicon sources. rf power is shown to be transferred into the plasma via two channels. The first is realized by weakly damping helicon waves which are excited directly by the azimuthal parts of the antenna and penetrate into the bulk plasma. Electrostatic waves arising at a plasma edge owing to the linear mode conversion form the second channel. Strongly damped electrostatic waves can reach a plasma core at low magnetic fields only, while at high fields they deposit energy at the periphery of the plasma column. A principal fraction of the rf power is transferred into the plasma via the electrostatic channel, and so the power input turns out to be volume at low magnetic fields and surface at high fields. An enhanced volume input is possible at high fields in special anti-resonance regimes when the excitation of the electrostatic wave is suppressed. The effective collision frequency is introduced to describe effective damping of helicon waves arising due to their conversion into electrostatic waves. The results obtained permit explanation of both measured field profiles and the high absorption efficiency in helicon sources.

Polysilicon gate etching in high density plasmas. III. X-ray photoelectron spectroscopy investigation of sidewall passivation of silicon trenches using an oxide hard mask

The characteristics of poly-Si trench etching in high density plasma processes was studied by x-ray photoelectron spectroscopy. Poly-Si films on SiO2-covered Si(100) substrates were masked with a 200-nm-thick oxide hard mask. The 200 mm wafers were then etched downstream using a helicon high density plasma source and a chlorine-based gas chemistry. After etching, samples were transferred under ultrahigh vacuum to a surface analysis chamber equipped with an x-ray photoelectron spectrometer. Regular arrays of trenches were used to determine the photoelectron signals originating from the tops, sidewalls, and bottoms of the features. A thin oxide film was found on the sides of the oxide masked poly-Si trenches. The origin of this film can be related to the sputtering and redeposition of oxide from the quartz tube of the helicon source located in the plasma generation region. A substantial amount of chlorine was present on the poly-Si sidewall of the features, whereas less chlorine was found on the oxide surfaces. The poly-Si sidewalls were covered by a small amount of oxygen.

Polysilicon gate etching in high density plasmas. II. X-ray photoelectron spectroscopy investigation of silicon trenches etched using a chlorine-based chemistry

We have used x-ray photoelectron spectroscopy (XPS) to study the chemical constituents present on the surfaces after etching of poly-Si features masked with photoresist. Poly-Si etch rates were determined using in situ HeNe laser ellipsometry. The wafers were etched in a low pressure, high density plasma, helicon source using a chlorine/helium gas mixture. The XPS analysis chamber is connected to the reactor chamber via an ultrahigh vacuum chamber allowing quasi in situ analysis of the etched samples. Electrostatic charging of insulating surfaces and geometric shadowing of photoelectrons by adjacent photoresist lines have been used to differentiate photoemission signals from the tops, sidewalls, and bottoms of the features. Blanket and photoresist masked poly-Si samples were etched without biasing the substrate. The poly-Si etch rate decreased as a function of etching time and finally stopped due to the deposition of an oxide film. When biasing the sample, etching occurred and anisotropic etching profiles could be obtained. A thin oxide film was found on the sidewalls of the photoresist and poly-Si after etching. The thickness of the oxide film increased with etching time. No silicon was found on top of the photoresist. The silicon oxide film deposited on the surfaces resulted from the sputtering of the quartz tube located in the plasma generation region. Substantial amounts of chlorine were found on the tops, sidewalls and bottoms of the features. The chlorine concentrations on the sidewalls of the features etched at high bias power were comparable to those found on the horizontal surfaces when a low bias power was applied. The chlorine concentrations on the tops and bottoms of the features increased with increasing bias power. A low coverage of oxygen was found on all surfaces. The oxygen concentration on the sidewalls of the features also increased with etching time. Less oxygen was found on the bottom of the poly-Si trenches. Small amounts of carbon were detected on the bottoms of the features. A substantial coverage of carbon was present on the poly-Si sidewalls and was about half that found on the tops and sidewalls of the photoresist. These results indicate that the etching of the photoresist and the erosion of the quartz tube may enhance anisotropic etching.

Experiments and modeling of a helicon source

We report on experimental results from the Lawrence Livermore National Laboratory Helicon Source plasma source and numerical models used to understand the observations. These include measurements of plasma density with Langmuir probes, and heating measurements with an electron energy analyzer to test transit-time heating. The models describe helicon wave excitation, and electron heating. The wave fields are calculated using the ANTENA code including collisional and electron Landau heating. Numerical and analytic models are used to describe transit-time heating of electrons and quantify the effect of collisions.

Particle contamination characterization in a helicon plasma etching tool

There is much current interest regarding the formation, transport, charging, and behavior of particulate contamination in high density plasma tools, as these tools are generally regarded as the future of plasma processing for the semiconductor industry due to the need to obtain greater anisotropy and faster process rates concurrently with reduced surface damage. Because of the low pressures in which these tools typically operate (<5 mTorr), the likelihood of homogeneous nucleation processes leading to particle contamination problems is low. Similarly, the effect of ion drag is also expected to be greater, possibly leading to reduced particle trapping effects. Yet, few laser light scatter studies have been performed in high density plasma tools. This study is the first in situ characterization of particle contamination in a plasma etching tool using a helicon source. It was performed during normal process conditions for poly-Si etching and also under intentionally altered process conditions designed to produce particles. The effect of a bipolar, electrostatic chuck on wafer particle deposition was also investigated. Results showed that under normal process conditions, few particles were deposited onto the wafer; those that were observed were attributed to thermal stress effects resulting in flaking of deposition films, probably on the quartz bell jar of the source. Results also suggest that the electrostatic chuck increased particle deposition when the clamping voltage was applied. No trapping was observed over the wafer or near the source. However, some trapped particles were observed below the wafer platform. These particles showed unusual motion, but probably have minimal effect on wafer contamination. The issues pertaining to particle contamination formation and transport in this high density plasma tool are discussed.

Modeling of profile effects for inductive helicon plasma sources

A computer code for modeling existing and new helicon sources for materials processing has been developed. The Nagoya type-III, helical, and Stix coil antennas have been modeled to study and examine plasma density and temperature profile effects on power absorption of a small fraction (n/sub fe//n/sub e//spl ap/5%) of fast electrons (T/sub e-fast//spl ap/40 eV) which provide ionization of the neutral gas in the experiment, and bulk (T/sub e-slow//spl ap/3 eV) electron distributions in an argon gas. The "ANTENA" computer code, originally written by McVey (1984) to study ion cyclotron waves, was modified and used to study and model helicon sources. A collisional model that includes radial density and temperature profiles was added to the code to study the effect of collisions on the heating mechanisms. The competing effects of collisional and Landau damping heating mechanisms have been investigated in detail, and results indicate that collisions play an important role in the plasma absorption profile at high densities (n/sub c//spl ges/10/sup 13/ cm/sup -3/). The radio frequency wave absorption profiles are sensitive to the plasma density and temperature profiles. The partial-turn helix antenna, that solely excites the m=+1 azimuthal mode, is found to be more efficient in coupling the power to an assumed plasma profile than the Nagoya type-III. The Stix coil is also found to be promising due to its on-axis peaking of the wave heating fields.

Polysilicon gate etching in high density plasmas. I. Process optimization using a chlorine-based chemistry

High density plasma etching processes using chlorine gas have been developed for 0.25 μm polysilicon etching. Polysilicon films on SiO2-covered 200 mm silicon wafers are masked with photoresist patterns, and then etched downstream using a high density helicon source made by Lucas Labs of Sunnyvale, CA. To enhance the anisotropic etching, ion bombardment is provided by radio frequency (rf) biasing of the sample. Polysilicon trenching can be suppressed by performing a two-step etching process: the first uses a high energy ion bombardment resulting in high etch rate (250 nm/min) and anisotropy and the second a lower ion energy. The second etching step shows a high selectivity (≳80) of Si over SiO2 and therefore ensures the complete removal of the polysilicon during the overetch time. Perfect anisotropic profiles can be obtained without any trenching or other undesirable anomalies. The optimum etch rate nonuniformity is 6.5% (3σ) when operating the source at maximum rf source power (2500 W), 2 mTorr reactor pressure, and adding 30 sccm of helium to a 50 sccm chlorine discharge. Polysilicon and gate oxide etch rates have been measured using a real time in situ HeNe ellipsometer. Etch rates for polysilicon and oxide increase as a function of rf bias power but show no significant pressure or rf source power dependence. After etching, the 200 mm wafers can be transferred (under high vacuum) to an ultra high vacuum analysis chamber equipped with a Fisons ESCALAB 220i x-ray photoelectron spectrometer.

Power required to maintain an electron in a discharge: Its use as a reference parameter in magnetized high frequency plasmas

Earlier works on discharges sustained by electromagnetic surface waves in absence of a magnetic field have revealed the central role played by the power balance per electron. This balance relation stated that, provided energy transport is negligible, the power θL lost by the electron on the average in collisions with heavy particles is exactly compensated under steady-state conditions by the power θA taken by the electron on the average from the high frequency (hf) field, their common value being the parameter θ. Then, because θL is to a first approximation the same in all hf discharges under given discharge conditions and power density, a simple discharge model valid for all hf plasmas was used. The present article is an extension of this approach to hf magnetized plasmas, using surface-wave plasma columns placed in an axially directed static magnetic field as a means of investigation. We observe that θ decreases monotonously when increasing the magnetic field intensity B0, showing no extremum at or close to the electron cyclotron resonance frequency match over the gas pressure range (5–100 mTorr) investigated. We show that θ is controlled either by classical ambipolar diffusion or anomalous diffusion, the actual diffusion regime depending on whether the novel parameter B0p (p is gas pressure) is small or large. Our measured θ values are further used to estimate the average electron density in helicon sources given the power density, showing fair agreement with the reported values.

An experimental study of breakdown in a pulsed Helicon plasma

Fast time-resolved measurements of ion and electron energy distributions in the first 200 mu s of a discharge in a 20 cm diameter glass Helicon source show that, 10 mu s after the RF is turned on, an intense burst of electrons is produced with average energy 200 eV and lasting about 1 mu s. This initial pulse of high-energy electrons can be explained in terms of resonant secondary electron multiplication (the multipactor effect) of electrons which are accelerated in the RF fields produced by the Helicon antenna. These electrons ionize the background gas very rapidly and a plasma with high electron temperature and hence high plasma potential can be created in about 1 mu s. Subsequently, further ionization occurs and, after 20 mu s, the plasma potential has fallen to about 200 V and the electron temperature to about 20 eV. The electron temperature continues falling and reaches a steady state value of 8 eV in about 30 mu s. The density and plasma potential reach steady state after about 70 mu s.

Resonances and anti-resonances of a plasma column in a helicon plasma source

We consider excitation of a plasma cavity in a range of parameters typical for helicon sources. The resonance and anti-resonance properties are shown to be intrinsic for a plasma column in this case, as well as for any bounded plasma system. If a column is well separated from a conductive wall, even small collisions result in the destruction of both resonances and helicon anti-resonances. The anti-resonance states with quenched Trivelpiece-Gould waves turn out to be stable against collisions and ensure enhanced rf power release in the bulk plasma.

Silicon roughness induced by plasma etching

A parametric study of single-crystal silicon roughness induced by an SF6 plasma has been carried out by means of atomic force microscopy. An helicon source (also called resonant inductive plasma etcher) has been used to study the relation between plasma parameters and subsequent surface damage. The surface damage has been examined in terms of height roughness analysis and in terms of spatial (lateral) extent of the surface roughness. The central result is that roughness scales with the ratio of the ion flux over the reactive neutral flux (J+/JF), showing the combined role of both ionic and neutral species. At low ion flux, the neutrals smooth the surface, while at higher ion flux, they propagate the ion-induced defects, allowing the roughness to be enhanced. Influences of other parameters such as exposure duration, ion energy, or substrate temperature have also been quantified. It is shown that the roughness growth is well described by an empirical law: rms∝(1/√E)(J+/JF)ηtβ, with η≊0.45 and β≊1 (rms is the root mean square of the roughness). In other respects, we analyze the data with a Fourier transform analysis. The main advantage is to minimize noise and to separate the magnitude of the roughness, the lateral correlation length on which the roughness is growing, and the behavior of short and long range roughness. The results are identical to the rms analysis, especially, the above scaling law. The time evolution of the lateral correlation length follows a scaling law (which is not accessible by means of the rms) leading to a fractal dimension of 2.67. Also is observed a variation of the short range roughness as a function of the substrate bias voltage. Consequence for further scaling down of integrated circuits is called to mind.

Helicon ion source for plasma processing

The results of theoretical and experimental investigations of dense plasma source resonantly excited by helicon wave are presented. Driven by a single-loop low-voltage (≂300 V) antenna, a source generates a plasma of density up to 5×1011 cm−3 at gas pressures 0.5–20 mTorr, operates at magnetic fields <100 Gs, and produces low-energy ion fluxes (<60 eV) with current densities up to 7 mA/cm2 and 3.5% nonuniformity within a diameter 10 cm. A high-efficiency helicon source originates in a special kind of ‘‘resonance’’ wave discharge, where a sharp increase of plasma resistance occurs near the helicon wave dispersion branches.

Tungsten etching in low-pressure SF6 plasma: Influence of the surface temperature

The influence of the surface temperature on tungsten etching in a SF6 plasma diffusing from a helicon source has been studied in detail. The surface temperature dependence of the etching kinetics has been analyzed. The influence of the other parameters such as oxygen content of the plasma and ion energy have also been considered. The etching mechanism depends on the temperature range: the etching is partly spontaneous at higher temperatures (T≳−20 °C), becomes a chemical sputtering type at intermediate temperature (−20 °C<T<−40 °C), and a chemically enhanced physical sputtering at lower temperature (T<−40 °C). Surface analyses by x-ray photoelectron spectroscopy provide an understanding of how the surface chemistry is modified with temperature.

Ion energy distribution functions in a multipole confined argon plasma diffusing from a 13.56-MHz helicon source

Ion energy distribution functions (IEDFs) are measured in a low-pressure multipole confined diffusion plasma produced by a 13.56-MHz helicon source and compared with the nonconfined case. The plasma confinement is efficient below 1-μbar pressure and induces a plasma density increase of a factor of 2 in the diffusion region. Uniform radial profiles over 15 cm diam are obtained at the bottom of the diffusion chamber. The presence of the magnetized wall has a major effect on the IEDFs at 0.5- and 1-μbar pressure conditions at distances over 10 cm from the source exit. The tail of the IEDFs, which corresponds to hot ions coming direct from the source to the bottom of the diffusion chamber, is not affected by the confinement. However, the peak of the IEDFs, which corresponds to the background density, is drastically increased due to the reflection of the ions on the magnetized wall. The ion collisionality, involving resonant charge exchange with neutrals, is enhanced by the confinement: the first reflection on the magnetized wall induces a longer path length in the diffusion area before recombination on the nonmagnetized bottom wall resulting in a better thermalization of the hot ions diverging from the source at very low pressures.

Resist etching kinetics and pattern transfer in a helicon plasma

We describe the first commercial etching system based on a helicon ion source and characterize it for etching organic films in an oxygen plasma. This single wafer etching system achieves a throughput of at least 30 fully processed 5 in. wafers per hour, which is comparable to the throughput of batch reactors. The etching chamber is equipped with a 13.56-MHz 2500-W helicon source, four low-field magnets to shape the plasma and support the helicon wave mode, and a 600-W radio-frequency chuck with He backside heat exchange and a temperature range from −50 to +125 °C. Etching rates and uniformity were measured on unpatterned resist-coated wafers, while trilayer resist patterns were used to study pattern transfer effects. Under nearly optimum conditions we obtain an etching rate of 1.31 μm/min, a throughput ≥30 wafers/h, nonuniformity ≤± 3%, selectivity ≊70 relative to SiO2 and nearly vertical etching profiles for all types of features having dimensions down to 0.25 μm. We describe the effect of process variables on etching rate, uniformity, selectivity, and etching profiles. We present a model based on multicomponent adsorption kinetics that fits the observed dependence of etching rate on the process variables. This etching system is being applied to several advanced lithographic schemes including dry-developed resists, bilayer lithography, and pattern transfer through organic antireflective coatings and planarizing layers.

Tungsten etching mechanisms in low-pressure SF6 plasma

A detailed study of tungsten low-pressure etching in a helicon source reactor was performed. In correlation with surface analyses (transmission electronic microscope, Rutherford backscattering, nuclear reaction analysis), a complete parametric study of the plasma and etching parameters versus the macroscopic parameters [gas pressure, radio frequency (rf) power, substrate bias voltage) has been carried out. Using a model developed by Hoffman and Heinrich for silicon etching (Proceedings of the 9th ISPC, Pugnochiuso, Italy, 1989, p. 1003), and taking into account the experimental results, it has been shown that tungsten etching mechanisms can be expressed by the formation and destruction of a low-density reactive top layer. Sputtering of this layer under argon-ion bombardment has been studied by optical emission spectroscopy. Consequently, the tungsten etch rate can be expressed as an analytical function of the macroscopic parameters such as gas pressure, rf power, and substrate bias voltage.

Electron acceleration in helicon sources

Evidence is given for wave acceleration of primary electrons in a 1 cm radius helicon plasma source. The second root of the dispersion relation is also seen at low B fields. Electrostatic confinement of primaries appears to be important at these small diameters.

Electron acceleration in helicon sources

Electron acceleration in helicon sources

Measurement and modeling of ion energy distribution functions in a low pressure argon plasma diffusing from a 13.56 MHz helicon source

Helicon sources coupled to diffusion chambers provide a new method for better controlling plasma processing. It is, however, important to understand the effect of diffusion on the plasma evolving from the source to the wafer situated in the chamber. The diffusion of a 300 W argon plasma from a 13.56 MHz (6 cm diam, 20 cm long) cylindrical magnetic field free source into a 30 cm diam, 20 cm long chamber has been experimentally and theoretically investigated. A mobile electrostatic energy analyzer was used to measure the electron temperature, plasma potential, plasma density, and ion energy distribution functions (EDFs) along the common axis of the source and chamber. An analytical model based on experimental results has been developed. Although there is a potential difference of around 40 V between the source and the substrate table, charge exchange collisions associated with the plasma expansion reduce the average energy of the ions impinging on a grounded probe at the bottom of the chamber to about 3 eV. At 1 μbar the number of collisions is too low to cool all of the ions accelerated in the field and a tail remains in the ion distribution function. Although not measured, hot neutrals should also be generated as a consequence of charge exchange collisions and will impinge on the substrate.

Plasma diffusion from a low pressure radio frequency source

Helicon sources are becoming a promising technique for providing high ion fluxes with controllable energy. It is still unclear how the plasma parameters evolve from the source to the diffusion chamber where the wafer is situated. To investigate this region electric probes have been used (Langmuir probe and energy analyzer) to measure electron temperature, plasma potential, plasma density, and ionic distribution functions along the axis of a symmetric Argon plasma diffusing freely away from a 6-cm-diam 20-cm-long, 13.56-MHz (Helicon) source operating at 10−6 bar and 300 W. The plasma density and potential are simply related by Boltzman’s relation. Although a potential difference of 37 V is measured between the source and the substrate, charge exchange collisions reduce the average energy of the ions arriving at the substrate to about 5 eV. This low energy should be advantageous in processing easily damaged thin films.