2026年01月15日 智能调机与DOE

What is Cpk? Process Capability Analysis for Semiconductor Manufacturing

Key Takeaway

Cpk measures process capability: Cpk ≥1.67 is excellent (0.6 PPM defect rate), Cpk <1.0 requires immediate improvement. AI-powered SPC tools automatically calculate Cpk from production data and trigger alerts when capability degrades.

In semiconductor manufacturing, Cpk (Process Capability Index) is the most critical metric for assessing process stability. Whether a fab is qualifying equipment or an equipment manufacturer is delivering a tool, the first question customers ask is almost always: “What is the Cpk?”

However, many engineers’ understanding of Cpk remains at the superficial level of “above 1.33 is acceptable.” This article provides a thorough explanation of Cpk — its meaning, calculation methods, and practical applications.

What Is Cpk?

Cpk stands for Process Capability Index. It measures the degree of stability of your process output relative to specification limits.

In simple terms: the higher the Cpk, the more stable your process and the less likely your product is to fall outside specification.

The Cpk formula is:

Cpk = min[(USL – mu) / (3 sigma), (mu – LSL) / (3 sigma)]

Where:

  • USL = Upper Specification Limit
  • LSL = Lower Specification Limit
  • mu = Process mean
  • sigma = Process standard deviation

How to Interpret Cpk

Different Cpk values correspond to different levels of process capability:

  • Cpk >= 1.67 — Excellent. Process capability is ample, with an extremely low defect rate (~0.6 PPM). This is the target for most semiconductor fabs.
  • 1.33 <= Cpk < 1.67 — Good. Process capability meets requirements, but there is room for optimization.
  • 1.0 <= Cpk < 1.33 — Marginal. Barely meets specification; monitoring and improvement are needed.
  • Cpk < 1.0 — Insufficient. There is a high risk of out-of-spec product; immediate improvement is required.

What Is the Difference Between Cpk and Cp?

Cp measures only the process “spread” (i.e., data dispersion) without considering whether the mean is off-center.

Cpk accounts for both spread and centering. If the mean is perfectly centered on the specification target, Cp = Cpk; if the mean is shifted, Cpk will be lower than Cp.

In actual production, Cpk is more meaningful than Cp because it reflects the true process capability.

What About Ppk?

Ppk (Process Performance Index) uses the same formula as Cpk, but differs in how the standard deviation is calculated:

  • Cpk uses the within-subgroup standard deviation, reflecting short-term capability
  • Ppk uses the overall standard deviation, reflecting long-term performance

Generally, Ppk <= Cpk. If the gap between the two is large, it indicates the presence of special-cause variation in the process (such as equipment drift).

Typical Applications of Cpk in Semiconductor Manufacturing

In semiconductor manufacturing, Cpk is widely used for:

  • Equipment qualification: Customers require that critical parameters meet Cpk >= 1.33 or 1.67 at tool delivery
  • Process monitoring: Continuous tracking of Cpk trends via SPC control charts
  • Yield analysis: Cpk is directly correlated to yield; Cpk = 1.0 corresponds to approximately 2,700 PPM defect rate
  • Equipment benchmarking: Cpk comparison across multiple tools of the same model to assess equipment consistency

What to Do When Cpk Falls Short

If Cpk is below the target value, there are typically two remediation strategies:

1. Reduce Spread (Improve Precision)

Investigate equipment stability, consumable condition, ambient temperature, and other factors to identify the root cause of data variability.

2. Correct Shift (Adjust the Mean)

If the mean has drifted from target, process parameters must be adjusted to re-center it. Traditionally this relied on manual engineering adjustments; today, R2R (Run-to-Run) automatic parameter tuning systems can perform this correction automatically.

In actual production lines, equipment drift is the primary driver of Cpk degradation. During operation, consumable wear, temperature fluctuations, and other factors cause process parameters to gradually deviate from target. If this drift can be monitored in real time and automatically compensated, Cpk can be maintained at a consistently high level.

Calculate Cpk Online

If you need to calculate Cpk quickly, you can use our free SPC / Cpk Online Analysis Tool. Simply paste your raw data for one-click Cp, Cpk, Pp, and Ppk analysis, with automatic control chart and histogram generation.

MST
MST Technical Team
Written by the engineering team at Moore Solution Technology (MST). Our team includes semiconductor process engineers, AI/ML researchers, and equipment automation specialists with 50+ years of combined experience in fabs across China, Singapore, Taiwan, and the US.

读完这篇,下一步可以很具体

获取一份产线 AI 评估,看看 NeuroBox E3200 / SECS/GEM 怎么接到您的设备。

把设备类型、当前数据接口、工艺目标或良率问题发给我们。工程团队会先判断适合 VM、R2R、Smart DOE、EIP 还是能源优化,再给出下一步建议。

  • 适合晶圆厂、设备商、工艺/设备/自动化团队
  • 可从 SECS/GEM、Modbus、PLC、CSV/历史数据开始
  • 不需要先提交机密 recipe 或客户图纸

我们只用这条信息做初步判断,不会要求您上传机密工艺数据。

Ready to get started?
NeuroBox E5200

Reduce trial wafers by 80% with AI-powered Smart DOE.

📅 预约演示 📞 021-58717229 contact@ai-mst.com
💬 微信联系我们
企业微信客服

扫码添加客服,直接对话