Session FT3: Plasma Deposition & Damage

In-situ Diagnostics for Plasma Enhanced ALD and CVD

Plasma process in industry has largely relayed on empirical data gathered from ex-situ measurements such as etch or deposition rate, composition, comformality, selectivity, critical dimensions, etc.  Empirical evidence derived from experiment and observation has helped wafer process from nodes to nodes for decades without raising significant issues. However, in the fast-moving pace to sub-nanometer nodes, such trial-and-error method has found itself inefficient and incompetent for timely, cost-effectively, and accurately delivering the desired process result. This is because the surfaces being processed change dynamically at the atomical scale such as during plasma enhanced atomic layer deposition and etch. Ex-situ measurements would not be able to gain the insight of real-time surface reactions and reaction products formation because it is destructive regarding surface-gas phase equilibrium.

For the ultimate sake of designing a process from first principles, in-situ plasma and surface diagnostics are the minimum requirements to rightfully “observe” the real-time and dynamic evolution of the surfaces modified by a plasma and gas molecule adsorption. Only by in-situ, surface chemistry can be measured and monitored without distributing the reaction, preserving the integrity of reaction and identifying transient intermediates.

In this talk, we will give some practical examples to demonstrate how in-situ plasma and surface diagnostics play a critical role in plasma process development, focusing on PEALD and PECVD. An integrated plasma-surface interaction measurement system has been built with in-situ plasma diagnostics to quantitatively access the species and their energy states of process radicals and ions along with in-situ metrologies to monitor surface reactions at each process steps dynamically. The in-situ measurement is also interplayed with first principles and micro kinetics simulations to elucidate how the type and energy of radical and ion species reacts with different surfaces sites, which in turn, generates thin films of different properties.   

Development of Yttrium Oxide Film Deposition Method using Low-temperature Microwave Excited Atmospheric Pressure Plasma Jet

Due to the continued scale-down of integrated circuits, contaminant particles on the wafer due to process chamber wall erosion by plasma-wall interaction cause significant damage, resulting in defective semiconductor device [1]. Therefore, advanced ceramics are widely used as plasma-facing materials including Yttrium Oxide (Y₂O₃) due to their high resistance and chemical stability [2]. And the mechanism of interaction between plasma and Y₂O₃ in deposition process using an organic solvent is not completely understood. In semiconductor manufacturing etching processes, corrosive gas plasma is more harshly rushing the components inside of the chamber. To deal with this issue, it is necessary to utilize a Y₂O₃ film with a thick and high density to meet the requirements of the complex structure inside the chamber for this application. In addition, cost-effective methods are crucial in semiconductor manufacturing.

    We have been developing the cost-efficient plasma-enhanced chemical vapor deposition method for Y₂O₃ film deposition by microwave excited atmospheric pressure plasma jet (MW-APPJ), introducing an organic precursor solution mixing with working gas. The deposited films by MW-APPJ have investigated for their properties and performance under dependence of temperature, gas flow rate, and distance in order to suit industrial applications.

    Preliminary we could succeed to obtain the Y₂O₃ film with a deposition rate of ~0.8 um/min. The observed results, which include optical emission spectroscopy observation, film property analysis, mechanism of deposition, and future plans, are discussed in detail.   

EUV-induced plasma: PIC modelling of a transient plasma fluxes

Extreme Ultra Violet (EUV; 13.5 nm) is used to image patterns onto wafers in high-end semiconductor manufacturing. The EUV photons have sufficient energy (92 eV) to ionize the hydrogen gas inside the lithography tools. This EUV-induced plasma has various positive and negative effects, ranging from material etching and degradation, to electrostatic particle deflection. These effects are complicated by the pulsation of the EUV light; it comes in short pulses: 100 ns of light each 20 µs (50 kHz). Therefore, the EUV-induced plasma is never in steady state, but always transient shifting between bursts of high-energy electrons (up to 78 eV) and decaying afterglow. We investigate this plasma and its effects using a dedicated 3D Particle-In-Cell (PIC) model.

In this contribution, we focus on the ion flux to a pellicle inside a lithography tool. A pellicle is a very thin freestanding membrane (centimeters in size, but only tens of nanometers thick), which protects the photomask from particulates. It is extremely thin to minimize EUV loss by absorption, but this makes it vulnerable for plasma degradation. Hence, it is crucial to know the ion fluxes to this pellicle both to predict and improve its lifetime. Our PIC modelling results show fluxes of high energetic ion for ~500 ns in between decaying low energy flux for 20 µs. The resulting ion energy spectrum is used as input for material choices and testing strategies that will enable semiconductor manufacturing for the upcoming <3 nm nodes.   

Understanding and controlling the role of ions during plasma-enhanced ALD and ALE

As semiconductor devices continue to push the boundaries of traditional process technologies, atomic-scale processing methods are increasingly vital. Among these methods, atomic layer deposition (ALD) has emerged as a critical technique for creating ultrathin films on the demanding 3D geometries involved. Concurrently, atomic layer etching (ALE) processes are being integrated into state-of-the-art semiconductor manufacturing flows. ALE typically employs plasma-based techniques to achieve anisotropic etching, while ALD frequently incorporates plasma enhancement to reach the required results.

This contribution focuses on the significance of ions in plasma-enhanced ALD and ALE. Most prominently, it addresses the ion-surface interaction in plasma-enhanced ALD, an aspect that has not been widely considered to date. Topics that will be discussed are: (1) the role of low energy ions (< 30 eV) on the ALD process for, e.g., grounded substrates; (2) the tuning of the ion energy (up to 300 eV) during plasma-enhanced by rf substrate biasing and the beneficial and detrimental effects that can be achieved by highly energetic ions; and (3) the implementation of tailored-waveform biasing for reaching very narrow ion energy distributions (full-width half-maximum < 10 eV) and hence very accurately controlled surface reactions during ALD and ALE.   

Evaluation of Interaction Between Substrate and Nanoparticles Deposited by Plasma Chemical Vapor Deposition

In plasma nanotechnology, the deposition of nanoparticles on a substrate is an important research topic, together with nanoparticle synthesis. So far, we have revealed the growth mechanism of nanoparticles in plasma chemical vapor deposition (CVD) to achieve size and transport control, while there is little information on how the nanoparticles deposited on the substrate are attached to the substrate. Here we analyzed the adhesion of nanoparticles deposited on the substrate to the substrate using a nanoindenter. In our experiments, we employed carbon nanoparticles (CNPs). We have successfully controlled the size of nanoparticles by plasma CVD and studied their transport [1, 2]. The nanoparticles were generated using CH₄+Ar plasma CVD. The size of CNPs was controlled by the gas flow. In this experiment, we fabricated CNPs of 20 nm in size. We deposited CNPs on a Si substrate on which 150 nm of hydrogenated amorphous carbon (a-C:H) film was deposited. The area covered by the deposited CNPs on the substrate surface was 10.9% of the substrate area. We measured the load-displacement curves of the substrate with and without CNP deposition were evaluated using a Berkovich indenter. The indentation depth was set to 15 nm to obtain information on the a-C:H film without the influence of Si. The hold time at maximum load was set to 1 s. For the sample without CNPs, the curves showed a typical a-C:H thin film. The Young's modulus and hardness were 200 Ga and 26 GPa, respectively. For the sample with CNPs, the maximum load was about one-fifth of that without CNPs and the modulus and hardness were 50 GPa and 3 GPa, respectively. The differences in the curves are considered to be information that includes the mechanical properties of the nanoparticles as well as changes in the contact conditions of CNPs on the substrate during the indentation. Details will be discussed at the meeting.

[1] S.H. Hwang et al, Processes 9 (2021) 2. [2] S.H. Hwang et al., Diam. Relat. Mater. 109 (2020) 108050.   

A Study on the Hydrogen Plasma induced material damage for EUV lithography Components and Material evaluation

In the ultra-fine semiconductor fabrication, the lithography process confronts a major change from deep ultraviolet (DUV) to extreme ultraviolet (EUV), introducing new challenges. Particularly, damage to components used in EUV system such as multi-layer mirrors, reticles, and pellicles within the lithography equipment due to EUV-induced hydrogen (H₂) plasma is a critical issue that directly impacts process yield and equipment lifespan. To resolve these issues, it is imperative to establish an environment which is similar to EUV-induced plasma and develop a method to assess the resulting damage. Accordingly, we developed an evaluation method of material damage by hydrogen radicals and ions in the inductively coupled plasma with H₂. In these system, the electron density was 5<!--[if gte msEquation 12]> style='font-size:11.0pt;mso-bidi-font-size:10.0pt;line-height:160%;font-family:
"Cambria Math",serif;mso-bidi-font-family:"Times New Roman";color:windowtext'> m:val="roman"/>× 10⁸ to 3.5<!--[if gte msEquation 12]> style='font-size:11.0pt;mso-bidi-font-size:10.0pt;line-height:160%;font-family:
"Cambria Math",serif;mso-bidi-font-family:"Times New Roman";color:windowtext'> m:val="roman"/>× 10¹⁰ cm^-3, the electron temperature was 1 to 4 eV, and the ion energy was several to tens of eV, which has similar environment of the EUV-induced hydrogen plasma. The damage of the Mo₂C, one of the possible candidate for EUV-pellicle material, samples was quantitatively analyzed by measuring the pore area fraction and chemical characteristics after plasma exposure under various H₂ plasma conditions such as electron density, gas pressure, and exposure time.