Overview: QSense QCM-D analysis in Chemical Mechanical Planarization (CMP)

To read an in-depth case study about QSense analysis of CMP slurry additive and abrasive interaction with a ceramic surface, click here.

Chemical Mechanical Planarization (CMP), sometimes referred to as Chemical Mechanical Polishing, relies on several nanoscale surface interaction processes, and so does a successful CMP result. Regardless of whether your ambition is developing and customizing high performing slurries, improving contamination control, or achieving reliable CMP protocols, it all starts with understanding and controlling the surface – molecule interactions involved.

In this overview, we present how QSense^® QCM-D technology can be used in molecule – surface interaction analysis and optimization of CMP-related processes. We also demonstrate what information QSense measurements offer.

This overview covers analysis of:

• additive – surface interaction
• post-CMP cleaning and slurry residue removal
• surface etching behavior based on slurry compositions

Characterize and optimize the chemical part of your CMP process

The chemical part of CMP represents a key aspect of the method. It is therefore important that this is optimized for the end use. Both in the development of new slurries and protocols, as well as in the improvement of existing ones, it is important to understand the chemistry and the surface interactions involved. With QSense technology, you can study surface processes related to, for example, slurry additives and their interaction with the surface. Some questions QSense can help you answer are:

• How does this slurry additive interact with the target surface?
• How does the effect of the additive vary with concentration? Which concentration is optimal?
• How does the slurry etch this surface material?
• What is the additive removal rate?
• How efficiently does this post CMP cleaner remove the residue?

Analyze material uptake and removal to get the full picture of the surface interactions

QSense is a surface sensitive technology that gives you a new set of eyes in the study of surface interaction processes. It is based on the Quartz Crystal Microbalance with Dissipation monitoring technology (QCM-D) which allows you to study nanoscale mass, thickness, and viscoelastic changes of your surface material or coating. The QCM-D method is time-resolved, which means that you can follow the interaction processes in real-time. Interactions and events that can be analyzed are, for example: adsorption, desorption, molecular binding, and structural changes such as swelling and crosslinking.

The layer structure provides unique insights

The QSense QCM-D senses hydrated mass, i.e. the mass of both the molecules adhering to the surface and the coupled solvent. This makes it an excellent technology to analyze systems where the layer conformation (i.e. structure) and degree of hydration varies with the measurement conditions.

Accurately simulate CMP process conditions

Surface interaction processes are highly dependent on the conditions where they take place. Therefore, when working with a real-life system, it is important to mimic these conditions in the experimental set up. QSense allows you to vary all the key parameters in your process. Running QSense analysis at relevant conditions, the surface interaction behavior of the additive, additive mixtures or slurries can be characterized and optimized.

What does QSense analysis of a CMP process look like?

Below we show examples of what typical data could look like in the different cases. We also discuss what information you could extract from the respective measurement and what questions you could typically answer.

1. A general example – monitor mass uptake and mass loss

The concept of the analysis method is straightforward. In brief, QCM-D technology monitors mass change as a function of time, i.e. it all comes down to monitoring mass uptake or mass loss from the surface when this is exposed to various samples and chemicals. A typical analysis could, for example, be that for molecular adsorption and desorption rates as well as the adsorbed amount.

The QCM-D measurement is based on monitoring of the two parameters, Δf and ΔD, Fig. 1, where Δf reflects the mass change and ΔD reflects the viscoelasticity of the layer. Combined, the captured raw data can be used to quantify the layer mass, Fig. 1, as well as the layer thickness, and viscoelastic properties (not shown).

Figure 1. Schematic illustration (not to scale) of (top panel) (A) molecule adsorption and (B) desorption, characterized by QSense QCM-D (middle panel).

The Δf and ΔD data reflect mass change and layer softness, respectively. As indicated by the grey arrows in the schematic graphs, the time-resolved data makes it possible to follow the adsorption and desorption processes, how fast they are, and how much material is added to and lost from the surface.

The amount adsorbed to, and desorbed from, the surface can also be analyzed via quantification of the time-resolved layer mass (bottom panel) or layer thickness as well as viscoelastic properties (not shown).

As all key conditions can be varied, the amount of information that can be extracted from this seemingly simple analysis is immense.

2. Characterize additive – surface interaction

Via the time resolved QSense measurement, it is possible to monitor how additives, additive mixtures, and abrasives interact with the surface of interest. The adsorption and desorption processes can be assessed in terms of how fast they are, and how much material that adsorbs to, or desorbs from, the surface as a function of the chemicals, surface, and solvent conditions, Fig. 2. As QSense technology also senses the mechanical properties of the surface-adhering layer, it is possible to analyze the arrangement of the adsorbed molecules via quantification of layer thickness and viscoelastic properties.

Figure 2. Schematic illustration (not to scale) of how the performance and behavior of different additives can be compared using QSense analysis. The example shows four different measurements where the parameter settings, in blue text, are varied relative to the measurement conditions in (A). In (B), the additive concentration is reduced, in (C) the surface material is changed, and in (D) a different additive is used. The time-resolved mass curves reveal the additive – surface interaction rate and the additive net mass uptake in the respective measurement.

3. Analyze post-CMP cleaning and slurry residue removal

QSense technology can be used to explore how to best remove post-CMP residue. In this case, the cleaning efficiency and residue removal is analyzed by looking at the mass loss, i.e. the mass removal rate and total amount removed. To run such an analysis, the chemical to be analyzed is first added to build up a layer at the surface, Fig. 2. Next, the surface adhering layer is exposed to the cleaning agent of interest and the mass removal is monitored.

Figure 3. Schematic illustration (not to scale) of how post CMP cleaning and residue removal can be analyzed with QSense QCM-D. The surface of interest is exposed to the slurry, or slurry additive, to build up a layer at the surface. Thereafter, the cleaning agent is introduced and the removal rate and removed amount is monitored as a function of time. Via comparison of the removal rates and net mass loss, the efficiency of different cleaning agents can be compared. For example, in the schematic illustration here shown, cleaning agent I would be both faster and more efficient than cleaning agent II

Use QSense QCM-D to:

• Investigate the removal of residues by post CMP cleaners
• Quantify the removal rates
• Quantify the amount removed from, or remaining at, the surface after the wash step
• Compare the efficiency of different cleaning agents
• Identify the optimal concentration of the cleaning agent

4. Assessment of surface etching behavior for different slurry formulas

In the previous section, we talked about QSense analysis of residue removal via monitoring the mass loss from the surface. Another process that would show up as surface mass loss is that of etching. Exposing the sensor surface to the slurry, or etching solution of interest, the etching process can be monitored as a function of time, Fig. 4. In the schematic example, the etch behavior of two different slurries, A and B, are compared. Slurry A is etching faster than Slurry B, and the net material loss is higher for slurry A when the experiment ends.

Figure 4. Schematic illustration (not to scale) of what QSense etching measurements of two different slurries could look like. The measurements start with bare surfaces in a non-etching solution to establish stable baselines. Thereafter, the slurries, or etching solutions to be studied, are introduced and the sensor material loss is monitored as a function of time. Finally, the measurements end with a rinse step to remove the slurries, and the net material loss can be quantified and compared.

Use QCM-D for example to:

• Analyze how the slurry etches different materials
• Investigate how the additive concentration affects the etching behavior
• Explore if the etching process is complex, i.e. if the slurry component first adsorbs then etches
• Assess how fast the etching stops when the slurry is removed

Concluding remarks

QSense analysis can be used to understand the chemical part of the CMP process. For example, it can be used to:

• Measure selectivity by varying the surface material of the CMP process
• Measure post CMP cleaning – determine if extra steps are needed to remove residue
• Measure point of use additives and their effect on etching or passivation
• Analyze slurry chemical removal or etch rates