InsightFab
Knowledge Base/Supply Chain Traceability: Designing a Traceability System from Raw Materials to Finished Products
Supply Chain Quality6 min read

Supply Chain Traceability: Designing a Traceability System from Raw Materials to Finished Products

This article explores the critical importance of traceability in the semiconductor industry, using an example where a sudden drop in wafer yield led to millions of scrapped wafers due due to a chemical issue, with multiple suppliers and used batches complicating root cause identification. It emphasizes why a proactive traceability system is essential, as reactive investigations are often insufficient to prevent significant losses.

"Once that batch of material was released, we knew we'd be working overnight to find answers."

Remember that incident late last year? The wafer fab's yield suddenly plummeted, dropping directly from a stable 99.8% to 97.5%, meaning millions of wafers were scrapped. The boss's face was ashen, and the production line supervisor was as anxious as an ant on a hot pan. Our engineers were either reviewing log files or meeting with equipment suppliers from morning till night. After two days of investigation, we discovered it was an issue with one batch of chemicals. But the problem was, the same batch of chemicals had been supplied by more than one vendor, and several drums within the factory had already been opened and used. So, how do you pinpoint which drum, from which supplier, caused the problem in that situation? Just thinking about it gives you a headache. This is precisely why "traceability" is a matter of life and death in a semiconductor factory.

What's the Root of the Problem? Simply Put, "Not Knowing Where to Start"

You might think traceability is just about coding items and recording them? It's not that simple. Semiconductor processes are complex; from silicon wafers, various chemicals, and gases, to photomasks and targets, an error in any single step can cause yields to drop irreversibly. Previously, we would only start investigating when a problem occurred, but by then, it was often too late to stop the bleeding.

Therefore, the key is that we need a system that can clearly understand the "ins and outs" of every batch. From identifying the upstream raw material supplier, when it entered the factory, and its batch number, to which machine it entered, which batch of product was used, and what finished product was ultimately produced. All information follows the product like DNA. This way, when a product has an issue, such as an abnormal yield for a certain batch, we can immediately trace back to which raw material, which machine, or even which parameter setting was incorrect.

How is it Actually Done? Implementing a "Comprehensive Tracking System"

To achieve true traceability, the core is "data linkage." We integrate Manufacturing Execution Systems (MES) and Enterprise Resource Planning (ERP) systems to bind data from all stages together.

  1. Raw Material Entry – Assigning an "ID Card": Each batch of raw materials entering the factory must have its batch number scanned and linked to supplier information, incoming date, chemical analysis reports, etc. This is like giving it a unique ID.
  2. Machine Parameters and Batch Number Linkage: When raw materials enter a machine for production, the MES automatically records which batch number of raw material is used, on which machine, at what time, and with what parameters processing occurs. For example, when a specific etching machine processes wafers with batch number P1234, it records an etching time of 60 seconds and a power of 500W.
  3. Finished Product and Process History Linkage: Each wafer ultimately produced will have a unique wafer number. This number will be linked to all machines it passed through during the entire process, all raw material batch numbers used, and the process parameters at that time.

In other words, when a batch of finished products has a Cpk of only 1.08, far below the target 1.33, and the DPMO skyrockets to 6210 ppm, we can use the system to directly retrieve all raw material batches used and all machine parameters involved for this batch of finished products. This efficiency is a world of difference compared to manually checking tables in the past.

Most Common Pitfalls: Human Error and System Silos

Honestly, the biggest pitfall in implementing a traceability system isn't technical issues, but "people." The most egregious case I encountered was a field operator, in a hurry, pouring raw materials in directly without scanning them. Consequently, when that batch of material had issues, the traceability system showed "raw material batch number unknown," rendering all efforts futile. Therefore, the system design must enforce scanning before proceeding to the next step, and regular audits are necessary.

Another major pitfall is "system silos." System A manages raw materials, System B manages processes, and System C manages finished product shipments, but they are not interconnected, and data cannot be exchanged. In essence, "everyone does their own thing," and the result is still manual cross-referencing. Therefore, it is crucial to ensure data connectivity between systems from the initial planning stage.

One Thing You Can Do Today

Start by implementing batch number scanning and system linkage for the most critical raw materials in your department.

Want to try it yourself?

Every tool mentioned in this article is available on InsightFab — just upload a CSV to analyze.

Go to Tools

More in this category

Part Approval Process (PAP): How to Substitute PPAP

This article highlights a critical challenge faced by a semiconductor manufacturer: a supplier's component Cpk report of 1.08, a quality level fundamentally unacceptable for their industry, creating a significant impasse. It reveals the substantial divergence between the stringent part validation standards of the semiconductor sector and the traditional automotive PPAP, particularly concerning yield requirements.

Writing Purchasing Quality Specifications (PQS): Clearly Stating Your Requirements to Suppliers

This article uses a jarring anecdote of a production line machine's 1.08 CPK and 6210 DPMO report, coupled with a supplier's "Your specifications said so" defense, to underscore a crucial point: quality failures often stem from inadequately defined requirements. It emphasizes the critical importance of a clear, quantitative Purchasing Quality Specification (PQS) in preventing issues caused by vague standards and lazy template reuse.

Establishment of a Supplier Evaluation System: From Candidate to Qualified Supplier

Many companies struggle with suppliers that are highly recommended but ultimately fail to meet quality standards. This article shares a hard-earned experience where a batch of new wafers in their factory suffered a drastic drop in yield, with alarmingly low Cpk values, highlighting the critical need for a robust supplier evaluation system. It emphasizes that every detail, from equipment and process stability to personnel quality, must be rigorously assessed to avoid seemingly promising but ultimately costly "landmine" suppliers.