Presentation for NCCAVS – Plasma Applications Group (formerly ...

avsusergroups.org

Presentation for NCCAVS – Plasma Applications Group (formerly ...

Presentation for NCCAVSPlasma

Applications Group (formerly PEUG)

“Advanced Plasma Treatments to Improve 4

Semiconductor Device Packaging Applications:

Die Attach, Wire Bonding, Mold & Encapsulation,

and Flip Chip Underfill”

- - -

Oct. 8, 2009

1 | 08 October 2009 |


A Picture’s Worth 1,000 Words…

Untreated ASPA Substrate:

Droplet test indicates low wettability

2 | 08 October 2009 |

Substrate After Plasma Treatment:

Droplet test indicates high wettability

(correlates to both high surface energy

and increased bondability)


4 Common Plasma Applications for ASPA:

• Die Attach

• Wire Bonding

• Mold & Encapsulation

• Flip Chip Underfill (FCUF)

Reliable Wire Bonding enabled

by Plasma Treatment

Cohesive Mold Failure (Desired)

thanks to Plasma Treatment

Fast, Void-free Underfilling

after Plasma Treatment

3 | 08 October 2009 |


3 Common Types of Plasma Treatments

Hydrogen Plasma

Chemical Process

Oxide Removal

H. H .

Oxygen Plasma

Chemical Process

Organic Removal

H .

O

.

Argon Plasma

Physical Process

Sputtering

O

. +

Ar

Ar +

O

H

H

O

O

C

4 | 08 October 2009 |


5 | 08 October 2009 |

Plasma Before Die Attach


Plasma Before Die Attach

Plasma Treatment to:

• Activate surface of die attach pad for improved die bonding

• Remove contamination from die attach pad

• Benefits:

• Increase die-to-substrate bond strength

• Increase die-to-substrate bond reliability

• Reduce scrap/re-work

• Reduce costs

6 | 08 October 2009 |


Die Attach Adhesion Experiment

• Objective: Maximize die attach adhesion (shear testing)

• Test Set-up

• QFN devices

• Copper leadframes

• Silver-filled epoxy die attach (oven-cured)

• 15 leadframes plasma treated

• 15 leadfames untreated (control)

• Commercially avail. Plasma System

• Argon only

• Shear test 1 die per leadframe (chosen at random)

• Considerations

• Maximize cleaning & surface activation rates (for high throughput)

• No damage to leadframe wire bond pads

7 | 08 October 2009 |


Relative Shear Strength

Die Attach Adhesion Results (Shear Testing)

7

6

5

4

3

2

1

0

Die Shear Test: Untreated vs. Optimized Plasma

Failure Mode: Substrate-Epoxy Interface Failure Mode: Die-Epoxy Interface

Untreated

Plasma

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Plasma Treated Average: 5.29

Sample #

Untreated Average: 3.36

8 | 08 October 2009 |


Conclusions for Die Attach

Plasma Improve Die Attach:

• Improved die adhesion to substrate

• Shear testing confirms failure mode now at die

surface

• Optimized treatment rates

• Throughput matches die attach system

• Further testing by customer confirmed: No damage to

wire bond pads on substrate

(Important, will be discussed in following sections)

9 | 08 October 2009 |


10 | 08 October 2009 |

Plasma Before Wire Bonding


Plasma Before Wire Bonding

Plasma Treatment to:

• Remove Organic, Metal, Halogen, and Oxide contamination from

bond pads (both die & substrate)

• Benefits

• Increase wire bond (pull, shear) strength

• Increase wire bond yields (Cpk)

• Increase wire bond reliability

• Reduce scrap/re-work

• Reduce costs

Wire Bond Attach Rate ~100%

(both Ball & Wedge) thanks

to Plasma Treatment

Courtesy ASM, C. Vath

11 | 08 October 2009 |


Plasma Before Wire Bonding

• Wire Bonding Issues

• Fine Pitch Wire Bonding (thinner wires, more bonds, smaller area)

• Pad Contamination

• Bond Pad Metallization Issues

• Aluminum, Copper

• Nickel & Palladium

• Thin Film Gold

• Changing Substrate Technology

12 | 08 October 2009 |


Plasma Before Wire Bonding

Let’s examine one specific issue for Wire Bonding…

• Bond Pad Contamination by Organics

Wire

Intermetallic

Compound

Contamination

Pad

30 to 300

Angstroms

13 | 08 October 2009 |


Contamination Sources

(Causes reduced wire bond reliability & yield)

• Bleed of silver-filled epoxy (i.e. die attach step)

• Thermal, Snap Cure processes

• Resin bleed

• Organic contamination from other sources

(Unfortunately, many sources in a Fab)

14 | 08 October 2009 |


Epoxy Bleed - Plasma Cleaning Experiment

• Quad Flat No-Lead (QFN) Package

• Cu frame with Ni/Pd/Au

• 32 lead/die; 576 die/strip

• Silver-filled epoxy die attach

• 2 cases: Oven-cure, snap-cure

• 25 mm gold wire

• Commercially avail. Plasma System

• Argon & Oxygen mixture

Courtesy ASM, C. Vath

15 | 08 October 2009 |


Plasma Treatment to Remove Contamination

Plasma Generated

Reactive Species

Successful

Wire Bond

Active Surface

16 | 08 October 2009 |


17 | 08 October 2009 |

• Optimized Plasma

• Mean Pull Strength >10 grams; CpK 2.07

• No Plasma (control)

• Mean Pull Strength 3.89 grams; CpK 0.03


18 | 08 October 2009 |


Conclusions for This Contamination

Plasma can effectively remove epoxy bleed out

contamination

• Argon-Oxygen mixture is both sputtering & oxidizing the

contamination away

• Removal of contamination results in both higher

reliability & yield

• As measured by wire bond pull strength testing & CpK

calculations

Plasma Improves Wire Bonding

• Thermal Cure process

>250% improvement in wire bond pull strength

• Snap Cure process

>25% improvement in wire bond pull strength

19 | 08 October 2009 |


Plasma Before Wire Bonding

Let’s examine another issue for Wire Bonding…

• Bond Pad Metallization Issues (a.k.a. Bond Pad

Contamination by Metal Oxide)

Wire

Intermetallic

Compound

Contamination

Pad

30 to 300

Angstroms

20 | 08 October 2009 |


Oxide on Pad - Cleaning Experiment

• Common bond pad metalization: Nanometers of gold (Au) on

palladium (Pd) on nickel (Ni)

• Problem: Nickel migrates through gold, forms NiOx on surface

NiOx contamination causes wire bonding problems

• Solution: Plasma clean to remove NiOx

• However: Top layer of metallization (i.e. thin gold) can be

sensitive to sputtering

• Without expertise, possible to remove entire layer of gold in a single

plasma cycle!

• Challenge: Find a plasma treatment to clean the NiOx w/o

removing a significant amount of gold

21 | 08 October 2009 |


Oxide on Pad - Plasma Cleaning Experiment

• Quad Flat No-Lead (QFN) Package

• Cu frame with Ni/Pd/Au (25nm)

• 32 lead/die; 576 die/strip

• Silver-filled epoxy die attach

• Oven-cure only

• 25 mm gold wire

• Commercially avail. gas plasma system

• Argon only

• 25 years of plasma know-how

Courtesy ASM, C. Vath

22 | 08 October 2009 |


Plasma Treatment to Remove Contamination

Plasma Generated

Reactive Species

Successful

Wire Bond

Active Surface

23 | 08 October 2009 |


Optimized Plasma vs. No Plasma

24 | 08 October 2009 |


Optimized Plasma Is Required

Sample

Conditions

Mean Pull

Strength

CpK

No Plasma 3.72 grams 0.07

NiOx remains on pads

Under 4.67 grams 0.35

Treated

Optimized 8.52 grams 2.15

Over

Treated

25 | 08 October 2009 |

4.82 grams 0.45

Thin gold layer sputtered away


Conclusions for Bond Pad Metallization Issues

• Gas plasma system can effectively

remove NiOx contamination

• Argon-only plasma sputters the contamination away

• Removal of contamination results in both

higher reliability & yield

• As measured by wire bond pull strength testing &

CpK calculations

• Stay within plasma process window to

maximize performance improvement

• Under-treat: NiOx remains

• Over-treat: Thin gold is sputtered away

26 | 08 October 2009 |


27 | 08 October 2009 |

Plasma Before Mold & Encapsulation


Plasma Before Mold & Encapsulation

Plasma Treatment to:

• Increase surface energy

• Enhance flow characteristics

• Reduce wire bond damage (sweep)

• Remove organic contamination

• Modify surface of material (“roughen”)

• Benefits

• Increase mold strength (peel, shear)

• Increase mold yield (Cpk)

• Reduce scrap/re-work

• Reduce costs

Adhesive Mold Failure

(Not Desired)

w/o Plasma Treatment

Cohesive Mold Failure

(Desired)

w/ Plasma Treatment

28 | 08 October 2009 |


Mold & Encapsulation: Why Plasma

• Customer’s Objectives: Improve Yield and Reliability

• Maximize Mold Adhesion

• Minimize Delamination

• Reduce Mold Flash (Deflash)

• Reduce/Eliminate Wire Sweep

• Issues for Mold & Encapsulation Process

• Fine pitch wire bonding

• Contamination

• Pre-processing

• Mold flash

• Leadframe materials

• Substrate itself

• Plating

Un-molded Leadframe

29 | 08 October 2009 |

Molded Leadframe


Mold Adhesion Experiment

• Objective: Maximize Mold Adhesion for Metal

Leadframes

• Test Set-up

• QFN devices attached with silver-filled epoxy (oven cure)

• Copper leadframes

• EMC mold material

• 10 leadframes plasma treated

• 10 leadfames untreated (control)

• Note 4 molded areas per leadframe

• Considerations

• Maximize cleaning & surface activation rates (for high throughput)

• No damage to bonded die, wire bonds and leadframe

30 | 08 October 2009 |


Mold Adhesion Experiment

• Evaluation Method

• Post-mold pull test (clamp-type)

• Test 1 site (out of 4) per leadframe

• Random site for each leadframe

• Equipment

• Commercially avail. Plasma System

• ASM Pacific Technology Mold System

Leadframes being Plasma Treated 4

Strips / Cycle

31 | 08 October 2009 |


Relative Pull Strngth

Mold Adhesion Results

Mold Pull Test: Untreated vs. Optimized Plasma

12

10

8

6

4

2

0

Untreated

Plasma

1 2 3 4 5 6 7 8 9 10

Plasma Treated Average: 9.12

Sample #

Untreated Average: 1.31

32 | 08 October 2009 |

Why such a large improvement


Mold Adhesion Results

• Pull Test Failure Modes

• Without Plasma: Mold separated cleanly from leadframe

• Mold not sufficiently bonded (Adhesive Failure)

• With Optimized Plasma: Cohesive Failure

• Mold so strongly bonded that failure destroyed the sample

33 | 08 October 2009 |


Conclusions for Mold & Encapsulation

Plasma Improved Mold Adhesion

• Maximized mold adhesion

• Minimized delamination

• Optimized treatment rates

• Throughput matched mold machine

• Further testing by customer confirmed: No damage to

bonded die and wire bonds

34 | 08 October 2009 |


35 | 08 October 2009 |

Plasma Before Flip Chip Underfill (FCUF)


Plasma Before Flip Chip Underfill (FCUF)

Plasma Treatment to:

• Increase filet height

• Increase filet uniformity

• Improve adhesion

• Minimize voiding

• Increase wicking speed

• Benefits

• Increase underfill adhesion

(both peel & shear strength)

• Increase underfill yield (Cpk)

• Increase underfill reliability

• Reduce scrap/re-work

• Reduce costs

36 | 08 October 2009 |


Flip Chip Underfill (FCUF) Issues

• Small gap between die and substrate Slow flow

• Trend is: Bigger (L x W) die, more bumps, smaller gaps

• Multiple material interfaces Voids & cracks

• Device passivation

• Solder Balls/Bumps/Pillars

• Substrate material

• Filet Height Usually the higher, the better

• Filet Uniformity Want uniform filets on all sides

• Voids, gaps, cracks Need to eliminate

• Thermal performance Delamination over time

37 | 08 October 2009 |


Plasma to Improve FCUF Performance

• Customer’s Plasma Objectives

• Increase Filet Height Improve Reliability

• Maximize Filet Uniformity Improve Reliability

• Maximize Wicking Speed Increase Throughput

• Minimize voids, cracks, gaps Increase Yield, Improve reliability

• Improve adhesion improve thermal performance reduce

delamination over time

(Remember this image)

38 | 08 October 2009 |


Flip Chip Underfill Experiment

Plasma Recipe Variables (to optimize plasma treatment)

• Gases, Combination of Gasses

• Power

• Pressure

• Time

• System Configuration (H/W, S/W, etc.)

• 25 Years Plasma Expertise

Must get plasma to go between Flip Chip and Substrate

• Flip Chip Samples

• 11 mm x 11mm bumped die

• Polyimide-based final die passivation

• Organic (FR4-type) substrate

• Measureables

• Filet height

• Filet uniformity

• Wicking speed

39 | 08 October 2009 |


FCUF Results – Filet Height, Uniformity

• Filet Height

• Without Plasma:

• Dispense Side:


FCUF Results – Wicking Speed

• Time Required for Underfill to Flow Under the Die

• Without Plasma Treatment: 40 Seconds

• With Optimized Plasma Treatment: 13 Seconds

41 | 08 October 2009 |


Conclusions for Flip Chip Underfill

Plasma Improved FCUF Performance

• Improved Filet Height

• Improved Filet Uniformity

• Increased Wicking Speed

• Customer Benefits

• Improved Device Reliability

• Increased Throughput

• Increased Yield

Also:

• Reduces voids that lead to cracks ( Possible topic for next time!)

• Reduces temp-cycle delamination ( Possible topic for next time!)

• Reduces device failure over time ( Possible topic for next time!)

42 | 08 October 2009 |


Thank you!

Q&A Session

43 | 08 October 2009 |

More magazines by this user
Similar magazines