Previous Page  31 / 60 Next Page
Show Menu
Previous Page 31 / 60 Next Page
Page Background


Chip Scale Review September • October • 2019


Innovation in inspection and metrology for new

manufacturing technologies

By Scott Jewler

[SVXR, Inc.]

n 1984, KLA i nt roduced t he

2020 Wafer Inspector. With this

innovation, the wafer fabrication

i ndust r y began the t ransit ion f rom

the manual optical inspection of the

time to today’s artificial intelligence-

powe r ed fab -wide proce s s cont rol

with feedback and feedforward loops.

Automatic inspection and metrology

have since become cor nerstones of

wafer manufacturing with as many as

three out of ten process steps dedicated

to inline monitoring.

I n s e m i c o n d u c t o r p a c k a g i n g ,

au t oma t e d op t i c a l i n s p e c t ion h a s

b e e n w i d e l y d e p l oye d . Howe ve r,

because critical interconnect features

are frequently optically obstr ucted,

this technique is unable to provide

feedback on the performance of core

processes such as f lip-chip bonding.

As i nt e rcon nect densit y cont i nue s

t o i n c r e a s e , a n d a s i n c r e a s i n g l y

complex and multi-layered assembly

t e ch nolog ie s a r e w i del y a dop t e d ,

and as market tolerance of process

escapes and field failures diminishes,

it has become clear that transmissive

i n s p e c t i o n a n d m e t r o l o g y o f

semiconductor assembly processes is

needed to increase yields, lower costs,

and eliminate escapes.

Ac o u s t i c m i c r o s c o p y, m i c r o -

focu s X- r ay, mag ne t ic r e sona nc e ,

and i n f ra red ( IR) technolog ies a re

all finding their ways into packaging

f a c t o r ie s . Wh i le e a ch t e ch nolog y

serves a useful purpose, none offer the

combination of speed, sensitivity, and

resolution needed for inline continuous

process monitoring.

A newly developed high-resolution

automatic X-ray inspection (HR-AXI)

technology overcomes these limitations.

Combi ned with advanced ar tif icial

intelligence algorithms, 100% inline

inspection and metrology of critical

optically obstructed features can now

be accomplished.

Why use HR-AXI for advanced


Interconnect densit y in advanced

packag i ng is i nc rea si ng at a r apid

pace. 2.5D packages with hundreds of

thousands of solder joints are shipping

in volume. Complex system-in-package

(SiP) modules now include integrated

shielding, multiple active components,

finely spaced passive components, and

double-sided surface mount technology.

At the same time the interconnect

density has been increasing, the market

has grown much more unforgiving of

quality escapes and field failures. Social

media brings unrelenting attention to

new technology and blockbuster product

releases and rapidly amplifies reports of

any issues in performance or reliability.

A sma l l qu a nt it y of f ield fa i lu r e s

can redistribute billions of dollars of

market share.

Meanwhile, electronic devices take

on more and more responsibility for

human safety. Assisted and autonomous

driving depend on device reliability.

Wh i le r edu nd a ncy ca n r educe t he

risk of functional failure, the cost of

recalls and repairs only increases with

additional devices.

How does HR-AXI add value?

Ultimately, value determines if an

inspection or metrology step will be

added to the manufacturing process flow

or not. How does HR-AXI add value? HR-

AXI increases process tool utilization.

Typically, a sample is prepared after a

maintenance event or shift change and

moved to the lab for analysis. Depending

on the process and event, the process tool

may idle while the quality of the set-up

sample is confirmed. This confirmation

may be done by conventional X-ray or

cross sectioning, but either way, the

production capacity of the idled tool is

lost while waiting for the quality check

to be performed. HR-AXI eliminates the

need for offline process set up checks by

monitoring product quality immediately

and continuously on the production floor.

HR-AXI also improves cycle time.

Manufacturing cycle time drives agility

and efficiency in manufacturing. Shorter

cycle times mean less inventory carrying

cost, better responsiveness to demand

signal changes, and improved absorption

of overhead. By continuously monitoring

product quality at key process steps,

lots move more quickly through the

manufacturing line. Cycle time can be

optimized and savings captured.

Fu r t he r mor e, HR-AXI imp rove s

product quality and reliability. Statistical

process control (SPC) is effective at

minimizing variation due to common

causes of variation. Unfortunately, not all

causes of variation are common. Special

or assignable causes also exist and these

are not typically exposed by conventional

SPC techniques. As advanced packaging

process requirements get more restrictive,

it is also common to find cases where

proce s s capabi l it y is ma rg i na l t o,

or incapable of meeting, the design

spe ci f icat ion. A ma rg i na l proce s s

may be in control, but still produce

an unacceptable number of defective


Flip-chip contact non wets (CNW)


Figure 1

) are a particular concern.

Frequently occurring randomly, these

defects are very difficult to detect with

conventional techniques. Because they

often have assignable causes, SPC is


Figure 1:

Normal joint vs. contact non wet.