Skip to main content

Computer modelling and experimental validation of plasmas and plasma- surface interactions, for a deep insight in cryogenic etching

Periodic Reporting for period 1 - Cryoetch (Computer modelling and experimental validation of plasmas and plasma- surface interactions, for a deep insight in cryogenic etching)

Reporting period: 2016-06-08 to 2018-06-07

As the electronic feature dimensions in the semiconductor industry are continuously shrinking, porous materials are increasingly being used as inter-metal insulators to address the critical need for a low dielectric constant (low-k). However, the current state-of-the-art plasma etch recipes face challenges for etching porous material, i.e. plasma induced damage (PID), as radicals and ions can easily penetrate into the interconnected pores, causing severe damage.
Currently the most promising technique for fast low-k material etching with limited PID is cryogenic etching. By cooling the wafer to cryogenic temperatures in fluorocarbon based gas (e.g. C4F8 and C6F6) before plasma processing, this gas may condense in the pores as liquid, which can prevent the diffusion of radicals into the interconnected pores during the subsequent plasma etching. A fundamental understanding of the mechanisms is highly desired to further optimize the plasma etch processes.
Therefore, this project intended to obtain more fundamental insight in the underlying plasma behaviour during cryogenic etching, the surface reaction mechanisms, and the etch process with pore-stuffing (with condensed C4F8), by means of extensive modeling, validated by experimental diagnostics. This project provided a theoretical foundation and guidance for research and industrial applications of the plasma processing of porous material.
1.Mechanisms for plasma cryogenic etching of porous materials
Porous materials are commonly used in microelectronics, as they can meet the demand for continuously shrinking electronic feature dimensions. However, they are facing severe challenges in plasma etching, due to plasma induced damage. In this project, we studied both the plasma characteristics and surface processing during the etching of porous materials. The simulation results indicate that the interconnected porous network acts as a penetration channel for the radicals, which causes deep damage into the material stucture, and it can even destroy the dielectric spacing, leading to structure loss. When lowering the chuck temperature, the C4F8 gas used for pore stuffing gradually condenses inside the porous material and fills the pores. This pore filling can mitigate the PID at the sidewalls and also reduce the connectivity between the pores, i.e. the penetration channels of the radicals, which decreases the damage depth into the material. When the pore filling degree reaches 50%, the porous material starts to become well protected. However, since the filling first needs to be removed by ion sputtering before etching can proceed, the etch rate is also significantly reduced upon lowering the chuck temperature. Therefore, a chuck temperature of around -104 0C might be more beneficial, as it yields limited damage (in spite of the moderate filling degree of 15%), while still allowing significant etching. Comparison is made to experimental data to validate the simulation results.
This work is published in .

2.Propagation of a plasma streamer in pores of dielectric material
Understanding the micro-interaction between plasma species and porous material, e.g. how plasma streamers can propagate in the pores, and what is the minimum pore size to make this happen, is of crucial importance for plasma processing, as it defines the surface area exposed to the plasma species, and thus the ultimate performance of plasma processing.
We studied plasma streamer propagation and discharge enhancement inside porous material. We demonstrated that the Debye length is an important criterion for plasma penetration into the pores, i.e. plasma streamers can penetrate into the pores when their diameter is larger than the Debye length. Our simulations can predict the minimum pore diameter needed for plasma streamers to penetrate inside the pores for given operating conditions, and they provide a deeper understanding of the propagation mechanism of streamers inside these pores, for pore diameters ranging from nm to μm.
This work is published in .

3.Importance of surface charging during plasma streamer propagation in pores of dielectric material
Different porous materials will have different chemical effects, but in addition, they might also have different dielectric constants, which will affect surface charging, and thus the plasma behavior. We demonstrated that surface charging plays an important role in the streamer propagation and discharge enhancement inside pores of dielectric material, and in the plasma distribution along the dielectric surface, and this role greatly depends on the dielectric constant of the material. For εr < 50, surface charging causes the plasma to spread along the dielectric surface and inside the pores, leading to deeper plasma streamer penetration, while for εr > 50 or for metallic coatings, the discharge is more localized, due to very weak surface charging. In addition, at εr = 50, the significant surface charge density near the pore entrance causes a large potential drop at the sharp pore edges, which induces a strong electric field and results in most pronounced plasma enhancement near the pore entrance.
This work is published in .

4.Capacitive electrical asymmetry effect in an inductively coupled plasma reactor
The independent control of ion energy and ion flux is very important for the plasma processing of materials, as the ion energy controls the specific surface process, and the ion flux determines the processing throughput.
By using the chemistry reaction set of SF6/C4F8, we investigated for the first time ever the electrical asymmetry effect by applying multiple frequency power sources (13.56 MHz and 27.12 MHz) to a capacitively biased substrate electrode in a specific inductively coupled plasma reactor. On the one hand, by adjusting the phase angle θ between the multiple frequency power sources, an almost linear self-bias develops on the substrate electrode, and consequently the ion energy can be well modulated, while the ion flux stays constant within a large range of θ. On the other hand, the plasma density and ion flux can be significantly modulated by tuning the inductive power supply, while only inducing a small change in the self-bias. Independent control of self-bias/ion energy and ion flux can thus be realized in this specific inductively coupled plasma reactor.
This work is just accepted
This project provides us with a better understanding of the fundamental processes and mechanisms of plasma etching of porous material at cryogenic temperature, and the interactions between the plasma species and porous materials. The comprehensive understandings of the cryogenic etch process concluded from this project will facilitate the optimization of the etching of low-k porous material.