InsightFab
Knowledge Base/Semiconductor Yield Improvement Case Study: SPC Uncovers Hidden Process Drift
Semiconductor Process6 min read

Semiconductor Yield Improvement Case Study: SPC Uncovers Hidden Process Drift

This article details a semiconductor manufacturing case where an unexpected yield drop occurred despite all parameters appearing normal and quality checks passing. It reveals that the issue stemmed from an "invisible" process drift, gradually degrading yield, which could not be detected by merely observing individual data points or parameter limits. The piece highlights the necessity of using statistical process control (SPC) to identify these subtle, underlying process shifts before they lead to significant quality issues.

That Day, the CPK Report Came Out, and the Room Fell Silent for Three Seconds

I still remember several years ago when our machine suddenly produced a batch of products with abnormally low yield. At the time, the process parameters on the daily report all appeared green, indicating they were within specifications. The equipment engineers confidently guaranteed that "parameters hadn't changed," and quality control reported that "sampling inspections were all OK." Yet, strangely, wafers experienced significant yield drops during backend testing. The yield plummeted from an original 98.5% down to 96.2%. When the boss saw the report, his face turned livid. The atmosphere in the meeting room was extremely tense; everyone exchanged glances, and no one knew what the actual problem was.

Where Was the Problem? It Was That 'Invisible' Drift

To be frank, often when we monitor machine parameters, we only observe "point" data, such as whether the current temperature is 200°C or the pressure is 100 Torr. However, process parameters are inherently subject to "continuous" variations. Solely looking at individual data points makes it extremely difficult to detect that a parameter has gradually "drifted" away, much like the slow boiling of a frog. This "process drift" is hard to discern in its initial stages, but over time, its cumulative effect can severely impact yield. At this juncture, simply relying on parameter upper and lower limits to catch it is ineffective, as the parameter may still be within the specified control range.

So what should be done? This is where SPC (Statistical Process Control) becomes invaluable. SPC does not merely check if individual data points are out of specification; instead, it utilizes statistical charts to help us monitor the "trend" of process parameters, enabling us to detect those subtle, underlying issues proactively.

How Is It Done in Practice? Look at Trends, Not Just Points

At that time, we used SPC control charts to pinpoint the issue.

  1. Collect Sufficient Data: We retrieved all process parameter data from the past month, not just from the day when the yield was abnormal.
  2. Plot Control Charts: Plot these data points on a control chart. Control charts typically include a Center Line (CL), Upper Control Limit (UCL), and Lower Control Limit (LCL).
  3. Identify Abnormal Trends: In a normal process, data points should be randomly distributed above and below the center line. However, if any of the following conditions appear, it indicates potential process drift:

* Seven consecutive points falling on the same side of the center line (e.g., all high or all low).

* Six consecutive points showing a monotonic increasing or decreasing trend.

* Fourteen consecutive points alternating up and down.

That time, we discovered a certain critical etching parameter that exhibited a trend of "8 consecutive points above the center line" on the control chart. Although each individual point remained within the UCL and LCL, and the Cpk value was still 1.08, appearing not serious, the trend was unmistakably clear. In other words, while the parameter "wasn't out of specification," it had silently gravitated towards the upper limit, at which point the DPMO had already surged from 2000 to 6210.

The Most Common Trap: Over-Reliance on 'Not Out of Specification'

I'm telling you, the most common pitfall is placing too much trust in the statement "parameters are not out of specification." Often, machine parameters are indeed still within the control limits, and engineers consequently believe "there's no issue." However, process drift is akin to the slow boiling of a frog; by the time you discover it's out of specification, significant problems have typically already arisen.

I recall an instance where a newly recruited junior engineer observed a trend of seven consecutive low points on an SPC chart but dismissed it, thinking, "It's not below the lower limit." A few days later, the electrical performance of that product batch completely deviated, and the yield plummeted to below 90%. The furious boss threw the control chart in his face and exclaimed, "Didn't you see the trend?" This illustrates that understanding trends is far more crucial than merely looking at individual data points.

One Thing You Can Do Today

Go back and check your control charts to see if there's a trend of seven consecutive points on the same side!

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

Yield Engineering: From Die Yield to Wafer Yield

Ever wondered about wafer fab yield numbers but felt lost? This practical article introduces you to the world of wafer yield, using a vivid "pineapple field" analogy to clarify die yield, wafer yield, and what happens to defective chips. You'll learn how to pinpoint root causes from data, determining whether issues stem from equipment or processes, making crucial semiconductor knowledge accessible.

Yield Loss Classification: Random Defects vs. Systematic Defects

This article addresses the common factory challenge of unstable yield by differentiating between 'random defects' and 'systematic defects.' It provides a clear guide on how to identify the root cause of yield fluctuations, enabling more targeted problem-solving and significantly reducing time and cost spent on investigations.

Process Capability Assessment for Thin Film Deposition (PVD/CVD)

This article recounts a challenging experience where a newly promoted senior engineer faced intense pressure from management over a thin film deposition report showing a CPK of 1.08. It then provides an accessible introduction to thin film deposition, explaining common techniques like PVD and CVD. Readers will learn the critical importance of evaluating thin film process capability and the numerous intricacies involved in manufacturing these fundamental layers in chip production.