SNF Labmembers’ Meeting 2017snf.stanford.edu/about/LabmemberMtgSlides-Oct 2017.pdfbuilding completely separate fab spaces with separate owning rooms for wafers with Cu on them, for

SNF Labmembers’ Meeting 2017

Stanford Nanofabrication Facility

Agenda The Big Picture (Nick)

› About SNF and our mission› nano@Stanford and NNCI› ExFab and the future of fabrication› FY18 labmember survey

Operations› FY18 Financials and rate structure (Mary)› Your staff (Mary)› Safety (Carsen)› ExFab tools (Swaroop)› Contamination (Michelle)

Closing remarks and discussion2

3

The Big Picture

What is SNF today?

• A collection of shared fabrication spaces: the Cleanroom (9k ft2),

ExFab (3k ft2), and the MOCVD lab(1k ft2)

• One stop shop for micro- to nano-scale fabrication

• A diverse community. In FY17, there were 647 labmembers:

• The 484 Stanford researchers came from 24 departments

• 25% were external (139 from industry, 24 from other universities)

• > 55,000 user hours logged in FY 2017

5

1. Provide access to cutting-edge fabrication and characterization capabilities for the Stanford community.

• World-class staff to teach and maintain equipment• Acquire new capabilities to stay at leading-edge• Support and renew existing tools

2. Educate next-generation of leaders in fabrication, and sustain an active community of researchers and external stakeholders

• Training, mentorship, and technical advising• Run classes, workshops, and symposia• External partnerships and outreach

3. Operate this highly complex facility safely and fiscally responsibly• Develop and disseminate best practices methods• Work within service center accounting rules• Seek to be revenue balanced

Our Mission

Changes in 2017

6

CleanroomSPF

SNF Hangout Area

• Lab renovation. Light!

• Launch of Ex-Fab: Official date Feb. 2017

• New tool installations

• New faculty director

Who is SNF?

7

SNF Operations: Staffing

8

Guiding principles:

• Keeping the lab safe• Keeping aging equipment operational• Developing the technical staff to meet research needs, including ExFab• Streamline administrative and business operations

And: achieve this evolution with minimal disruption to ongoing research.

Big Picture: NNCI

- Awarded Sept. 15, 2015- Different from NNIN- 16 independent sites- Broader geographic scope- Broader range of capabilities- Coordinated by Georgia Tech (Oliver Brand, PI)- Cooperative agreement through 2020: possibility

of renewal through 2025

9

NNCI = National Nanotechnology Coordinated Infrastructure: http://www.nnci.net/

Stanford’s site is “nano@Stanford”

nano@stanford and the NNCI

What does this mean for Stanford? See Nanolabs.Stanford.edu

10

nano@Stanford is made up of:- SNF (Stanford Nanofabrication Facility)- SNSF (Stanford Nano Shared Facilities)- MAF (Microchemical Analysis Facility)- EMF (Environmental Measurements Facility)

Mission- Make advanced research resourcesavailable to all by:- Cultivating an external user base- Democratizing know-how through

education and outreach

11

nano@stanford Outreach and Education

Workshops:- MOCVD Workshop, August 2016- ALD Workshop, April 2017- Direct-Write Workshop, July 2016 and 2017- And more to come!! (Suggestions and ideas are

welcome! Contact Mary)

Training and Education:- SNF YouTube channel: snftrainingvideos- On-demand training: https://lagunita.stanford.edu- We want suggestions and volunteers! (Contact

Angela Hwang)

New Capabilities: ExFabWhat is ExFab? The “Experimental Fabrication Facility”

12

• An “experiment” to inform design and planning of future shared fabrication facilities at Stanford.

• Fabrication resources beyond electronics, beyond silicon, and beyond the Roadmap.

• Not “cleanroom” (though not “dirty”)

• More flexible, broader range of capabilities, lower barrier-to-entry.

• Companion to the System Prototyping Facility, next to the new SNF cube area

13

How is ExFab doing?

ExFab has been around since Jan 2015 and received ~$2M from SoE/EE in response to pledges from 31 faculty to support this program.

ExFab formally launched Feb 2016, with the qualification of the Heidelberg.

ExFab now accounts for 23% of total equipment use in SNF.

Equipment Hours/month charged(Rolling 12 month ave)

1. Evaluate the toolset quarterly to identify usage patterns and bottlenecks through May 2018• Target UFY19 NC funds for a significant refresh• ExFab tools are generally less costly to acquire and install

won’t require the entire budget

2. Identifying the tools• Survey the ExFab users for a 2nd ExFab pledge in late 2017• Allows faculty to vote with their dollars (anticipated usage)• Requires commitment from new SoE Dean

3. Continue to be resourceful in garnering resources for the lab• Reach out to equipment manufacturers that ExFab could be a demo site• Six month trial runs for tools – part of the RPA program (more later)

FY18 and Beyond: Capital Equipment Planning

Dieter Scientific Xplore Microcompounder& Microinjection molder

Alveole PRIMOMultiprotein Photopatterning

Raith 150-”THREE”

Nagase KOACH“One-minute” CleanroomLesker Sputter #2

for cleanroom processing

15

SNF Labmember Survey 2018

Why a survey?

- New leadership (SNF, SNSF – and the University)- Your input will inform conversations about the future

of experimental resources

What we want to know

- 8 months into ExFab, what works and what doesn’t?- What is SNF doing right? What is wrong?- How do we build community – within SNF, across

shared labs, and between departments and schools?- What new tools/capabilities should we explore?- What is the future of fabrication and processing at

Stanford?

Take the survey!

http://bit.ly/2wDe3kh

or

Operations

16

Prof. Nick MeloshSNF Faculty Director

Mary TangManaging Director

Equipment

Gary Sosa

Elmer Enriquez

Ted Berg

Finance & Admin

Jana Krchnava

Lori JohnsonFinance & Admin Manager

Process

Michelle Rincon

Usha RaghuramLead

Xiaoqing Xu

Uli Thumser

Maurice Stevens

Lab Operations

Mario Vilanova

Mike Dickey

Carsen KlineLab Manager

Mahnaz Mansourpour (80%)

Jim Haydon

Jim McVittie (40%)

Engineering

Ray Seymour

SNF Operations: Staffing

Currently have 18.2 FTEPlan for FY18:- 25% AFM Expert- 25% Ebeam Litho Expert

Swaroop Kommera

New to SNF!

Rolling 12 month average: Remains stable at $320K/month, which facilitates projections for FY18 (which runs roughly with the academic year: 9/1/17-8/31/18).

SNF Operations: Financials

19

SNF Operations: Rates

Proposed FY18 Rates:

Submitted, awaiting approval (expected by Nov.) Sep/Oct billing will reflect proposed rates.

Cleanroom: Same as in FY17

ExFab: Rates largely the same, reduced in some cases: industrial rates on select tools is 2X rather than 3X.

MOCVD: Reduced 20% from FY17.

SNF Operations: Detailed ExFab Rates

Proposed FY18 Rates:

ExFab: Rates largely the same, or reduced: industrial rates on select tools is 2X rather than 3X.

Major Labmember Projects and Programs

Student Helpers – Undergrads/grads who support training, monitors, projects. (Michelle)

On-demand training – Videos and courses. Funded by NNCI. (Angela Hwang)

Redefining Contamination classes – ProM Committee (Michelle)

E241 – ExFab Project class. Funded by University/SoE (Prof. Jon Fan)

EE310 – MOS Device Class (Usha)

Retiring legacy servers – Chromebooks, email lists, website (Carsen, Mary)

21

Safety first!

If you see something, say something Public announcement via [email protected] Private discussion with staff

Friendly reminders No hand carrying of chemicals Only approved chemicals allowed No powdered materials allowed in cleanroom Clean work surfaces when you’re done Leave no unknown/unlabeled materials

23

Evacuations

Thank you for your cooperation All evacuations were orderly and safe… And there were a lot of them

We’re working on it… Most alarms are outside our area, but we’re

working to resolve them Dusty gas vaults being cleaned Construction manager well aware of issues Buddy system for maintenance on gas lines

Allen/AllenX Fire Alarms 2016-2017

1/29/16 - Dirty smoke detector2/1/16 - Dirty smoke detector5/8/16 - Construction dust7/6/16 – Germane/hydrogen release during cylinder change7/9/16 - Construction dust1/1/17 - Sprinkler system faulty flow control5/30/17 - Hydrogen release during tool removal9/27/17 - Construction / fire panel wiring9/28/17 (2x) - dust in air handling duct

Aveole PRIMO Multi-protein patterning

• 375 nm laser• Resolution ~1.5um• Photocleavable PEG for direct patterning on glass• Can pattern SU8• Tiff or bitmaps accepted for patterning• No charge to use; demo through December

Nanoscribe Photonic GT 3D Printer

• Two-photon 3D printer • 250nm resolution in X,Y direction. 800-900nm resolution in Z • In the "maskless lithography" 2D writing mode, resolutions of

~100 nm can be achieved.

251,2 - Pictures from Nanoscribe presentation3,4,5,6 – Structures made with SNF Nanoscribe tool

1

23

45

6

From H. Chiamori

From W. Liang

Heidelberg MLA 150 Direct-Write Lithography system

26

• Direct-write on resist – no mask needed!• Substrates from < 1cm2 to 6”X6” in size• ~1 um resolution• Front-side overlay <0.5 um, Front-backside ~1 um• Fast writes (full coverage of 4” wafer in 8 min.)• 405 nm source for g-line and broadband resists• Gray scale for 2.5D/3D lithography• $35/hr for academics

Optomec Aerosol Jet 300 3D Surface Printer

• Prints almost any material in solution (Ag, Au, CNTs, Epoxy, SU8, proteins, live cells) with viscosity 1-1000 cp.

• Resolution to ~10 microns. • Focused stream enables writing onto non-uniform, highly topographical

substrates, up to 5 mm high.• Commercial application to flexible electronics, labeling on 3D substrates,

chip to package interconnects• Lots of possibilities for 3D writing of novel materials and substrates

System in 155A

(From the Optomec website)

(From the Optomec website)

Micromist Coater

• A research-grade, mini electrostatic spray tool• Blanket deposition of material• Advantages - little material waste and highly

conformal coating.• Materials as diverse as photoresist and chocolate

have been demonstrated on this.

28

Flip Chip Bonder

• For heterogeneous materials integration• Solder bump bonding, thermocompression, and

eutectic bonding of a variety of substrates

Fineplacer Lambda Flip Chip Bonder

Sapphire flip-chip thermocompression and eutectic bonding for dielectric laser accelerator. HUIYANG DENG, YU MIAO Mentors: Mark Zdeblick, Anthony Flannery, Usha Raghuram. Research Advisor: James S. Harris, Olav Solgaard (ENGR241 Fall ‘16)

Development of thermocompression and eutectic bond processes for pre-patterned substrates using the Finetech Lambda. Ki WookJung, Heungdong Kwon. Mentors: Usha Raghuram, Mark Zdeblick, Roozbeh Parsa, Anthony Flannery. Research Advisors: Kenneth E. Goodson, Mehdi Asheghi (ENGR241 Fall ‘16)

30

• Environmental control chamber (glovebox) • Ability to do AFM while applying bias to the device under test• SThM with electrical bias• Viscoelastic mapping for characterization of mechanical properties

Asylum AFM/Anasys

Coming soon: DISCO Backgrind

Precision removal of bulk substrate materials, including:- Silicon, compound semiconductors- Packaging resins- Copper-posts and other metals- Lithium Tantalate and lithium Niobate- Green ceramics and sapphireSubstrates from pieces up to 8”Reproducibility: <1.5 um across an 8” waferRoughness: <0.15 um

System is here, but installation will take ~2 months.

31

MANAGING METALS CONTAMINTION PROCESSING IN THE SNF: AN EVOLUTIONSNF STAFFERS 10/17

In the beginning, there were transistors…

Industry operates in a paranoid mindset since metal contamination kills yield

Metal films/flakes/ions at interface surfaces (at edges of etched structures, for example) create conductive paths where they are not wanted

Metal ions like to migrate around spreading electrons wherever they go, and depending on the metal and the material it is migrating through, wreak havoc on devices

Industry solution: Micro manage metals EVERYWHERE to mitigate risk- even building completely separate fab spaces with separate owning rooms for wafers with Cu on them, for example.

SNF modeled itself after industry as best we could within our budget:

Divide risk into categories based on metal type and level of potential impact

Dedicate equipment based on the risk category

Fast-forward to 2017New materials are being generated and evaluated constantly

These materials have major potential to impact research in many areas- from medicine to electronics to energy storage and beyond

Our SNF facility has the capability to enable researchers to explore!

The balance has shifted. There is much more demand for materials flexibility than contamination control.

Conundrum: How to co-exist.

StrategyShift attitudes around “paranoia” justifications that have been the semiconductor industry standard for decades

Introduce DEFENSIVE PROCESSING for those with strict contamination concerns

Create pathways (not necessarily dedicated tools) to allow research that requires research with strict metals contamination requirements to continue

Reduce hurdles for new materials work to move forward

Continuous evolution of contamination rules based on collaboration with researchers

Implementation:Define new categories:

CMOS Restricted

Flexible

Litho/Analytical

Tools will be distributed into categories based on functionality- i.e., there will be a CMOS restricted poly etch tool

Processes which need special attention will require PROM committee consultation to develop procedures. Decisions will be data-driven. If no data exists, it will be collected.

There ARE processes which need special attention. One example is clean metal etch. There is currently no “standard process” to go to a clean tool after a metal etch (the clean metal etcher died). Each case will be evaluated by the PROM committee, policies and tools will evolve from “market data” evaluated in PROM.

How does someone quantify contamination?Industry standard: Use VPD-ICPMS

Uses vapor phase HF (VPD) to decompose the native oxide and any contaminants on the wafer surface

The decomposition products are collected with a droplet of solution that is scanned (moved) across the wafer surface

That solution is evaluated by ICP-Mass Spec to identify contaminants.

Example: LixOy on Fiji3?Background: Fiji3 is a tool that is dedicated to non-conductive film processing only. It is very lightly used as of late, so there was a proposal to use it for LixOy generation. Would that be safe for our users who are growing gate oxides?

Plan:

1. Control: Place cleaned bare si-wafer face down on sample plate, load into chamber for 10 mins.

2. Run LixOy processes.

3. Place cleaned bare si-wafer face down on sample plate, load into chamber for 10 mins.

4. “Clean” the chamber: Run 100 cycles of Al2O3 as diffusion barrier to coat all surfaces.

5. Place cleaned bare si- wafer face down on sample plate, load into chamber for 10 mins.

ResultsLithium contamination clearly visible after ALD process

Post chamber-coat wafer did NOT show baseline Li levels- other metals look pretty good

Ca, K, Na often related to human handling- make sure to have control wafers!

Same tweezers used for post Li ALD wafer, did that cause contamination?

Conclusion: Users who are interested in gate oxide growth need to evaluate risk of Li contamination. If necessary, additional testing or chamber cleaning will be performed.

Litho/ Analyti

cal

Visual aid:

CMOS Restricted

Flexible

PROM consultation and approval REQUIRED