Transforming IC Interconnection Manufacturing: A Process Functional Modeling Analysis

In the ever-evolving landscape of semiconductor manufacturing, optimizing IC interconnection processes remains a critical challenge. Through innovative functional modeling approaches, we can identify inefficiencies, streamline operations, and dramatically improve production outcomes. This article explores how Process Functional Modeling (PFM) reveals new opportunities for enhancing IC interconnection layer manufacturing.

Setting the Stage

At PRIZ Guru, we are actively exploring IC interconnection layers. Our project is available on the PRIZ Platform and published on PRIZ Hub: Optimizing IC Interconnection: A Functional Approach to Innovation (the project is in progress). We welcome your engagement and comments.

The initial step—System Functional Modeling (SFM) of the IC interconnection layer—has been completed and is detailed in our blog article: Optimizing IC Interconnection Layer Design Through Functional Modeling.

For this next phase, we analyze the manufacturing process of the IC interconnection layer. We’ve chosen a typical interconnection layer that exemplifies both standard functionality and manufacturing processes. Using the Process Functional Modeling (PFM) tool on the PRIZ Innovation Platform, we’ve created a comprehensive map and analysis of the complete process flow.

Understanding Process Functional Modeling (PFM)

The Process Functional Modeling (PFM) tool in PRIZ Guru enables engineers to:

• Create detailed process flow maps and analyze each operation’s function
• Identify which operations add value and which create problems
• Develop targeted improvements by identifying operations to optimize, simplify, or eliminate

Learn more about PFM through these resources:

• Process Functional Modeling Overview
• PFM with the PRIZ Platform
• PFM Webinar Recording
• PFM Tool Instructions

Analysis Subject: IC Interconnection Layers

IC interconnections link various components within a microchip through multiple layers. Each layer requires the same set of manufacturing operations to be performed sequentially.

The original structure is shown below:

The next layer we need to build is shown below:

To achieve this, we follow a specific sequence of operations, illustrated below:

The process consists of 13 operations (steps). Note that metrology operations—which measure but do not alter the structure—are not shown in the picture.

Process Functional Modeling: Application & Results

Using the PRIZ Innovation Platform’s PFM tool, we mapped out a process flow incorporating standard metrology operations. The flow appears below:

For each operation, we defined these key features:

• Operation type: Productive (creates permanent changes), Providing (creates temporary changes), Corrective (removes defects), and Metrology (measures parameters)
• Functional Model: System Functional Models for each operation

PRIZ Platform calculated:

• Functional rank (shown as blue bars), Problematic rank (shown as red bars), and Operation Efficiency (OE) score to measure benefits versus drawbacks
• Key process metrics including Productive Operations Effectiveness and Operation Types Breakdown, with specific recommendations for process improvements

Our PFM summary shows that just 31.3% of operations add value to the product—the remaining operations consume time and resources without direct benefit. The analysis pinpoints which operations require development and which should be simplified or eliminated.

The screenshot of the summary appears below:

Key Findings and Recommendations

Our PFM analysis yielded several important insights.

Summary of the Process Functional Modeling

1. Building one interconnection layer requires at least 16 operations, including 3 metrology operations.
2. The process achieves a 70% average effectiveness. However, only 30% of operations add direct value to the product. This suggests we should eliminate or simplify the remaining operations.

Recommended Process Improvements

1. Etch Stop is needed only above copper (Cu) to prevent Cu diffusion from the bottom metal line to the top ILD. Instead of using CVD to deposit the Etch Stop material, we propose a selective deposition process using PMMA with dissolved SiC or SiN. This liquid solution can be spun onto the wafer and easily removed from the ILD (SiO2) areas.
2. Photolithography – We currently develop exposed resist using wet etch equipment. Since plasma can remove exposed photoresist (PR), we can skip the development step and apply plasma etch immediately after exposure.
3. Wet cleaning after dry etch – By applying plasma etch directly after exposure, we can eliminate both PR development and costly wet etch operations, as plasma effectively removes the exposed PR.
4. Polish (CMP) – To streamline the polishing operation, we propose removing most copper at the electroplater stage right after electrodeposition. By reversing polarity—converting the wafer to an anode and applying a negative charge to the existing anode—we can dissolve the bulk of Cu, simplifying the subsequent polishing step.

Strategic Shift

Photolithography investments need reevaluation. Since it’s a providing operation that adds no direct value to the product, we should prioritize its elimination.

Conclusion

Remember: When we purchase a product or service, we’re not just paying for its inherent value—we’re also paying for the costs of the producer’s unresolved problems.

Process Functional Modeling is a powerful creative thinking tool that can revolutionize problem-solving. Master it, and you’ll gain the ability to transform the world.

Have questions or need help getting started? Contact us today—we’re here to help you unlock the power of Functional Modeling!

Have questions or need help getting started? 
Contact us today – we’re here to help you unlock the power of Process Functional Modeling!