Thin Film Daily Abbreviations

膜厚飄移 N nm 或 N%。Thickness drift 是薄膜部門最常開的 deviation 主題。

• Thickness drifted by 3.2 nm on chamber B over the past 50 wafers.
• Thk climbed 5% after the precursor bottle swap — investigating.

均勻度超標 / 改善 / 崩潰。WIWNU 通常以 1σ% 表示,超 spec 是 chamber issue 警報。

• WIWNU is out at 1.4% sigma — spec is 0.8% — hold the chamber.
• Uniformity collapsed on chamber A after the showerhead swap.

顆粒貢獻數上升 / 飆高 / 平穩。任何薄膜 tool 出來 wafer 上 particle 多了就是 yield drop 警號。

• Particle adder count spiked from 12 to 80 on tool TF-04 — pull chamber.
• Particle count settled back to baseline after the wet clean.

膜剝落 / 分層 / 裂開 / 起泡。Peeling 通常是應力或界面黏著問題 — 整片 wafer 報銷風險。

• Film peeled on the M3 metal at the wafer edge — escalate to FA.
• Nitride delaminated on the dummy wafer — stress too tensile.

折射率偏移。RI 變化反映膜成分變化,通常意味 stoichiometry 飄了。

• RI dropped from 2.00 to 1.95 — likely the nitride is becoming N-rich.
• Oxide RI shifted from 1.46 to 1.48 — moisture suspect.

應力翻轉 / 過張 / 過壓。Stress 失控會 wafer bow 或膜剝落 — 兩個都終結 yield。

• Stress flipped from compressive to tensile after the chamber temp drift.
• PECVD nitride stress is too tensile at 800 MPa — wafer bows beyond spec.

腔體養爐 / 對比驗證 / 乾洗 / 濕洗。薄膜 tool 每 N 片要 clean,每次 clean 後要養爐 — 重複工。

• Chamber B was wet-cleaned and seasoned with 25 dummy wafers before re-qualification.
• Chamber A and B matched within 1.5% thickness — accepted.

前驅物被換 / 用完 / 認證。ALD / CVD 前驅物純度直接決定膜質,新批次到貨要 qualify。

• Precursor lot was swapped on 03/15 — re-qualified with 5 monitor wafers.
• TMA precursor is depleted in bubbler A — switch to bubbler B.

階梯覆蓋率不佳 / 可接受 / 優秀。深寬比大時 PVD 一定差,要切到 CVD 或 ALD。

• PVD step coverage is poor (40%) in the deep vias — switch to CVD.
• ALD conformality is excellent — 98% in 30:1 aspect ratio trenches.

8D · thickness excursion

The thickness drift on lot lot_id is attributed to root_cause, based on evidence.

{批號} 的膜厚偏移歸因於 {根因},依據 {證據}。薄膜 8D D4 標準寫法。

• The thickness drift on lot Q3-218 is attributed to showerhead clogging on chamber B, based on the SPC trend and post-PM monitor data.▸ 聽念法
• The thickness drift on lot 1842 is attributed to precursor bubbler temperature drift, based on the temperature logs.▸ 聽念法

PIE / TF disposition

WIWNU on tool_chamber is N% sigma; recommend action until WIWNU recovers to spec.

{機台/腔體} 的片內均勻度為 {N}% sigma;建議 {動作} 直到回到規格內。

• WIWNU on chamber B is 1.4% sigma; recommend hold and re-tune showerhead spacing until WIWNU recovers to spec.▸ 聽念法
• WIWNU on chamber A is 1.0% sigma; recommend monitor and re-cal next PM until WIWNU recovers to 0.8%.▸ 聽念法

particle excursion SOP

Particle adder count exceeded N on tool; chamber pulled for clean_type; recovery target target.

{機台} 上顆粒貢獻數超過 {N};腔體做 {清潔類型};目標 {目標}。Particle 異常的 SOP 句。

• Particle adder count exceeded 50 on TF-04; chamber pulled for deep wet clean; recovery target < 10.▸ 聽念法
• Particle adder count exceeded 30 on TF-09; chamber pulled for plasma dry clean; recovery target < 5.▸ 聽念法

change control · FMEA refresh

Adding new_step requires re-running FMEA on scope, including failure_modes.

新增 {步驟} 需要對 {範圍} 重跑 FMEA,涵蓋 {失效模式}。薄膜換 precursor / 新材料時必做。

• Adding the new TaN ALD precursor requires re-running FMEA on the M1 barrier, including peeling, stress, and contamination modes.▸ 聽念法
• Adding the new PECVD nitride recipe requires re-running FMEA on the M3 dielectric, including stress flip, RI shift, and step coverage failure modes.▸ 聽念法

customer escalation · peeling

Peeling observed on layer; customer impact assessment: assessment; N wafers under containment.

{層別} 發現膜剝落;客戶衝擊評估:{結果};{N} 片受 containment。Peeling 是大事故。

• Peeling observed on M3 metal; customer impact assessment: pending FA. 240 wafers under containment, none shipped.▸ 聽念法
• Peeling observed on the passivation layer; customer impact assessment: high risk. 48 wafers under containment; 12 already shipped — recall flagged.▸ 聽念法

Email — Particle Spike on TF-04

Thin Film 工程師早上發給 OPS + PIE,particle count 早班飆高。

Subject: [Action Needed] Particle adder spike on TF-04 — chamber pulled

Hi all,

Heads up — particle adder count on TF-04 spiked from 12 to 80 starting at 02:30 this morning.

Containment:
- TF-04 pulled from production at 03:15. 18 wafers from the affected window are held at the next monitor station.
- WIP rerouted to TF-06 and TF-08; capacity OK through end of shift.

Initial investigation suggests showerhead contamination after the routine PM yesterday. Equipment confirmed the clean cycle skipped the final flush step due to a controller fault.

Plan:
- Deep wet clean tonight.
- Re-qualify with 25 monitor wafers tomorrow morning.
- Target recovery: particle count < 10 before going back to production.

I'll send the re-qual results by Friday EOD.

Thanks,
Thin Film ENG

相關縮寫:particle · PM · WIP

Conf Call — Thickness Drift Update

PIE 主持月度 thin film capability review,Thin Film 部門 lead 報告 1 分鐘。

Thanks. Quick recap on chamber B thickness drift.

We saw thickness drift on chamber B starting two weeks ago. Drift is +3.2 nm relative to baseline on the PECVD oxide recipe; spec is ±2 nm.

Root cause is identified: precursor bubbler temperature drifted by 1.5°C after the room HVAC adjustment on 03/05. The bubbler control was within its own tolerance, but the system-level temperature shift wasn't caught by SPC because it's a slow drift.

Containment action: bubbler temperature LUT recalibrated; 5 monitor wafers confirmed thickness back to target. SPC stable for 96 hours now.

Long-term: we're adding the room temperature as an SPC parameter and tightening the bubbler temperature alarm window.

Cpk recovered from 0.95 to 1.51 over the last week. Capable.

Open to questions.

相關縮寫:thickness · SPC · Cpk · PECVD

Deviation Report — Film Peeling (Section 6)

M3 metal peeling 事故的 deviation report Section 6,寫的是 root cause + 改善。

6.0 Root Cause — M3 Metal Peeling Event

6.1 Symptom
Peeling observed on the M3 aluminum layer at the wafer edge (3-5 mm from edge) on 8 wafers across 2 lots (Q3-241 and Q3-242). FA cross-section confirms delamination at the Ti barrier interface.

6.2 Investigation Path
- Chamber correlation: 100% of peeling wafers ran on PVD-03 chamber B during a 6-hour window.
- Time correlation: incident window correlates with a tank pressure excursion on the Ar supply line (logged 17:42, recovered 23:50).
- FA: Ti barrier shows incomplete coverage at the wafer edge (3 nm vs 8 nm target).

6.3 Root Cause Statement
The peeling is attributed to incomplete Ti barrier deposition on the wafer edge due to insufficient Ar supply pressure during the 6-hour excursion window, confirmed by chamber correlation, gas supply logs, and FA cross-section.

6.4 Containment
- All wafers from PVD-03 chamber B during the affected window held (48 wafers).
- 100% edge inspection on remaining WIP that ran the same recipe.
- Tank pressure SPC alarm threshold tightened from ±10% to ±3%.

6.5 Long-Term Action
- New interlock: chamber pauses deposition if Ar supply pressure falls outside ±5% for more than 60 seconds.
- FMEA updated to include "gas supply pressure excursion" as a high-severity failure mode for barrier deposition.

相關縮寫:peeling · FA · SPC · FMEA · PVD

讀完了?來測驗看看記住了沒。