BS Poly Removal Defect Reduction Removal of edge cluster defects by improving recipe and hardware for backside polysilicon wet etch process

BS Poly Removal Defect Reduction Removal of edge cluster defects by improving recipe and hardware for backside polysilicon wet etch process

Malta, NY

Hong Zhai, Colin Weidner, Dhiman Bhattacharyya, Norberto DeOliveria, and Vincent Sih

TOC

1) Mission: repurposing an existing tool 2) Challenge: gap and performance 3) Approach and methodology 4) Current Result & status 5) Concluding remarks

2

Table of contents:

3

PC: Particle Count ER: Etch Rate

PC performance GOTO vs POR

4

GOTO POR PC

1.0 .75 .5 .25 0

Date-wafer

Half Year Later …

5

POR GOTO

Date-wafer

PC

1.0 .75 .5 .25 0

GOTO PC Qual Weekend Abnormality (A good way) Highlights 1. Failed to meet the target > half year. New base line proposed

2. We found 2 quals passed spec 3. Possible root cause needed to be correlated & discovered.

• Assumption base line shift needed to be explained.

Date-wafer

PC

1.0 .75 .5 .25 0

New Proposed Base Line

POR

Weekend

Carefully review this chart

7

Date-wafer

PC

1.0 .75 .5 .25 0

Stage of starting manufacture Qualification

Deep dive

Date-wafer

PC

1.0 .75 .5 .25 0

Re-Categorize the previous busy chart: (1) All wafer check one by one for history: 88 past runs wfrs.

(2) Red boxed: all new Bare Si wfr.

(3) Black box from one tool family.

(4) Non-boxed from any other tools in chart.

A few runs with controlled parameters

10

Controlled Run Controlled ReRun

Date-wafer

PC

1.0 .75 .5 .25 0

(1)Addition Pre Step. (2)Pre Defect measurement. (3)Process Step. (4)POST Defect measurement. (5)Additional Steps, like SEM etc

New Step

11

Two weeks watch after new plan implementation

12

Date-wafer-unit

PC

1.0 .75 .5 .25 0

SUMMARY REVIEW

What have done •Goto POR •DOI reduction

What to learn •FAB is BIG & sophisticated. •Improvement are all-facet. •Urgency of BIG DATA sorting.

What to proceed •GAPS remain. •Technologies distributions.

Current ER Performance .

Current PC Performance (GOTO only) .

POR

GO

TO

Current PC Performance (POR) .

ER

1.0 .5 0 .5 -1

Date-wafer-unit

Date-wafer-unit

PC 1.0 .75 .5 .25 0

1.0 .75 .5 .25 0

Edge cluster as Pin signature observed!

14

Check out this slide again! Two weeks watch after new plan implementation

Date-wafer-unit

PC

1.0 .75 .5 .25 0

An typical example

15

Date-wafer-unit

PC

1.0 .75 .5 .25 0

Actions taken and still continue…

16

Process • Spin speed. • Optimizing opening/closing the A and B chunk pins.

1) Exploring possible water droplets 2) Pins rubbing the wafer?

• A few different things: • - Tool family sequence compare. • - Adjust open/close sequences. • - Perform rinse segmentation. Hardware • Handling and pocket size.

To try: - Adjusting pocket size. - Perform Chuck A/B sequence

Recipe • Current (Rev 12) vs Rev 14 (150% increase spin speed) vs Rev 15 (A-B Dry) options. Pocket size:

• Water is pooling? • 150% increased spin speed test. Chuck Pin Open/Close • Chuck Pin Open/Close vs A/B switching times. • Two chambers showed promising data (reduced pin signature) with the pocket size at optimizing counts. • Change another set of MPCs to different pocket size. • Data that we gathered from these two pocket size.

Path forward:

(1)Pin Signature. GAP to close. (2)Correlation Pin signature with Prod. (3)Yield gain?

Concluding Remarks:

18

Summary: • Perform: POR Goto

Process step improve • Pin Signature: Pocket Size

adjustment need address

Thank you!