• No results found

Peak Sensor Systems e-diagnostics Initiative

N/A
N/A
Protected

Academic year: 2021

Share "Peak Sensor Systems e-diagnostics Initiative"

Copied!
17
0
0

Loading.... (view fulltext now)

Full text

(1)

Peak Sensor Systems

e-Diagnostics Initiative

Adding e-Diagnostic

Capability to ProPak Technology

Pam Ward / Peak Sensor Systems

[email protected], 505.342.1170, 106

(2)

ProPak Plasma Monitoring System

Current Features

:

– Fault library of known faults – Internal e-mail and pager

notification on Fault

– Historical data collection – Broadband OES technology

– Remote access via desktop tools – Dual methods for endpoint

– Playback tools for analysis

Pending Release: PlasmaTrax

– State of Health Data Base – Statistical analysis (PCA) – Return from clean

Features Under Development

e-Diagnostics

Notification internally and to OES vendor and OEM provider

Integration of Tool State on Board Sensors

– Data reduction technology – Fingerprinting

– Automatic algorithm development

(3)

Building Value

Endpoint Fault Detection State of Health

OEM Tool Parameter Archival

Fault Detection and State of Health Analysis OES in conjunction with Tool State Data

e-Diagnostic Communication of Tool Fault, Process Fault and Sensor fault to Quality Partners

Peak is building value by building on past successes. Adding nested layers of capability allows Peak to provide customized plasma monitoring solutions

(4)

Peak’s Historic Connectivity

Equipment Interface Engineer Desktop Engineer Desktop Engineer Desktop FAB Network Advanced Communications (Ethernet TCP/IP) DATA (SECS/GEM) Etch Cluster Tool

Pr o Pak OES (Fiber-Optics) e-mail / paging Comm SW SECS

(5)

Peak’s Vision of e-Diagnostics

ProPak monitors plasma process:

• All desired recipes and steps for endpoint

• Known fault conditions as they exist in a fault library

• Unknown fault conditions based on process trend excursion detection

ProPak monitors sensor condition

• Fiber condition, alignment • Tape backup status

• Spectrometer status • Communication status

ProPak monitors tool condition

• Critical tool parameters collected, archived, evaluated as part of process monitoring

• Validates tool parameters with OES data for more detailed fault analysis and diagnosis

• ProPak provides critical information regarding tool health using degradation analysis based on trend analysis

ProPak communicates between Quality Assurance Partners

(6)

e-Diagnostics Process Flow

OES Plasma Process Monitoring and Tool State Data Collection Process On? Plasma Monitoring Enabled? Yes No EP Detected? EP Normal? No Yes Process Completed Without Faults? No Yes Log Data Is it a New Day? No Send Heartbeat Message to Vender, OES Partner, Etc. Yes Process Fault Tool Fault Sensor Fault Log Data Notify: APC System, Tool Controller, Appropriate Contacts Internal Network Notification

of Engineer or Maintenance Personnel via Internal

Network

Internet Notification of External Vendors, OEM, Other Authorized Users

(7)

Peak’s e-Diagnostics Connectivity Plan,

No Additional Network

AEC/APC control system such as Fabworks,

Brookside,etc. (May not be available)

Engineer Desktop Engineer Desktop Engineer Desktop FAB Network DATA (SECS/GEM) OES (Fiber-Optics) Peak Sensor Systems

LAM ProPak IPC Fire Wall Process Fault Fault Mediation (Ethernet TCP/IP)

Etch Cluster Tool

Tool Status Tool Faults Fault Mediation Tool Diagnostics Sensor Status Sensor Faults Fault Mediation Sensor Diagnostics

Lot, Slot, Recipe, Step Start, Step Stop, On Board Sensor Data

Fire Wall Fire Wall Internet 2-way VPN Communication 2-way VPN Communication

(8)

Peak’s e-Diagnostics Connectivity Plan,

Additional Network

(9)

The User’s View

User receive heartbeat message from ProPak? Yes No Quality Partners are alerted to heartbeat message failure User loads production into etch chamber Plasma processing begins ProPak determines if plasma monitoring enabled Yes No ProPak remains idle

Has process fault occurred?

Has process tool

fault occurred? Has sensor fault occurred? ProPak monitors for

endpoint and fault conditions

OES & Tool state data collected Yes No No Yes Yes No Was endpoint successfully detected? Yes Data and process

results logged Quality Partners notified, fault information logged with data No Quality Partners are notified of endpoint failure, data logged with failure details

Service Call initiated by OEM

or Sensor Vendor. Service Call Notice goes

to User

Is fault tool or sensor related?

Yes Process Fault Notice is

sent to User

No Plasma process halted, damage

(10)

Technical Demonstration

Detection of known fault conditions based upon OES

data merged with tool state data.

Communicate the occurrence of fault

– Process fault: Facility Engineer (SEMATECH, Austin) – Sensor fault: Sensor Vendor (Peak Sensor Systems,

Albuquerque)

– Tool fault: OEM Provider (LAM Research, Fremont)

Test Details

– 100% Silicon represented “normal” product

– 100% Photoresist represented a “no mask” process fault

– A detached and blocked fiber represented “Broken Fiber” sensor fault

(11)

Fault Details

Ratio (722-730) to (652-662) 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 5 10 15 20 25 30 35 40 Time (sec.)

Normal Wafer (No PR) 100% PR

Broken Fiber +25% Cl -25% Cl

652,662,722,730,1.33 -- 2st Test for Broken Fiber

722,730,652,662,0.33 -- 5th Test for Low Cl2 Flow 652,662,722,730,2.17 -- 4th Test for High Cl2 Flow

Note: these tests may catch other unpredicted problems since these bandpasses are sensitive to other parameters of a well behaved process. Light Leak

(12)

Fault Details

Ratio (360-415) / (505-525) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 5 10 15 20 25 30 35 40 Time (sec.)

Normal Wafer (No PR) 100% PR

Broken Fiber +25% Cl -25% Cl

360,415,505,525,0.75 -- 1st Test for 100% PR

(13)

Understanding Plasma Parameter Space

Plasmas reside inside of an operational space.

Plasmas are dynamic systems that strive to maintain a balance by compensating for slight drifts in gas flow, pressure, power and temperature.

Ideally, plasma parameters are

picked so that the plasma resides in the center of the parameter space, resulting in robust plasma systems tolerant to larger parameter shifts.

OES is not a quantitative tool, but rather qualitative. It can only

respond when parameters have shifted the plasma out of the space.

(14)

Peak’s Goals for e-Diagnostics

Unprecedented Customer Support

e-Diagnostic capability allows Peak Sensor Systems to

provide an unprecedented level of customer support

– OEM/Vendor partnerships build on e-Diagnostic capabilities all

collaborative trouble shooting via remote connectivity

• 24 Hour process surveillance

• Immediate response to trouble calls

– Remote connectivity allow OEM/Vendor partners to provide JIT service and response

• Predictive tool and sensor maintenance

• Instantaneous installation of software upgrades • Instantaneous service contract support

– With remote access to the tool sets in the field, OEM/Vendor partners can provide remote performance monitoring and tool operation

• Engineering support to develop fault detection methods, endpoint algorithms, etc.

(15)

What Went Wrong

Lessons Learned

Peak Sensor Systems became involved in the e-Diagnostic

program relatively late in the year (October, 2002)

The short time to develop and test was negatively impacted

by the holiday seasons, making it difficult to run repetitive

tests

The original tool demonstration was performed the week

before Christmas, impacting the number of attendees to

the demonstration

Demonstrations were run with PR blanks and Si blanks,

therefore the tests had no real-world applicability

Simulation of tool fault type was not as sensitive as we

(16)

What Went Right?

Peak and Lam received excellent support and active

collaboration from ISMT without which these successes

would have not been possible:

– Communications between the Sensor and Quality Partners, Peak, Lam and ISMT were quickly established.

– VPN access to the in Fab OmniPak system was simple and elegant once established.

– Remote access allowed Peak engineers to perform much of the required development on the Austin system from Peak’s Albuquerque office-Saving $$$ and Time while offering a higher level of customer support. – System configuration, data retrieval, and sensor software upgrades were

performed remotely without any issues.

– Tool parameter gathering and archival was successful.

– Communication of tool status (heartbeat messages, fault conditions) was delivered to all quality partners immediately upon occurrence.

(17)

The Future of e-Diagnostics at Peak

Where Peak wants to go from here:

– Diagnostics based on tool parameter deviations – Linking the Fabs via ProPak

– More detailed fault information – Incorporation of additional sensors

– Addition of predictive maintenance capability to ProPak

• Predictive cleans

• Impending tool failure

• Detailed diagnostics with “Just in Time” replacement part acquisition to maintain tool and sensor performance.

– Establish 24 hour Customer support centers to provide in fab support via electronic monitoring.

• EP development

• Fault algorithm development • Etc.

– Enhanced Security

• User identity and activity logs • Security keyed information

References

Related documents

Neutrophils express MMP-9 in response to direct M.tb infection or stimulation by M.tb -driven monocyte-dependent networks and in patients with CNS-TB.. First, we investigated

In addition, alliances with structural powers increase the chance of success of leadership strategies (de Águeda Corneloup & Mol, 2014, pp. Therefore, it is expected that

The major reported barriers for treatment were lacking financial means to afford care, fear of being perceived as weak for having mental health problems, fear of being perceived

— cognitive (scienti¿ c) — exposed parts of the two environments may be subject to scien- ti¿ c research and education both of the geological region as well as of the changes

I will now begin by considering the relevant and available literature in relation to those with refugee status and learning, as well as exploring the wider societal

Blacklick Creek and Codet Road were the planted sites and had lower species richness (12 and 13 species) and diversity (Hʼ=2.1 and 2.0), respectively. Eight tree species were

announcement date,

"Formulation of food grade nanostructured lipid carrier (NLC) for potential applications in medicinal-functional foods", Journal of Drug Delivery Science and Technology,