3D InCites Member Spotlight: Why the Front End Needs the Back End To Win in AI

We recorded live at IMAPS with Siemens, ACM Research, Shellback Semiconductor, DECA, Nordson Electronic Solutions, and VIEW Micro Metrology to explore how AI demand, chiplets, and panels are reshaping advanced packaging. We dig into 3D BLOX, thermal and test roadblocks, green chemistries, metrology at scale, and why the back end now leads innovation. Listen to learn about:

• The Siemens–ASE collaboration on 3D BLOX models and VIPACK workflows
• Interoperable YAML-based packaging definitions moving toward IEEE standard
• 3D stacking to cut picojoules per bit amid thermal and test limits
• Panel-level packaging economics, sizes, and lack of standards
• ACM Research updates in copper plating, bevel clean, frame clean, and compound deplating
• Batch spray versus single wafer trade-offs at Shellback Semiconductor
• HydrOzone green strip replacing legacy NMP in select flows
•The  DECA–SST deal for NVM chiplet package and SoC disaggregation
• Nordson Electronic Solutions' panel strategy, IntelliJet 1.1, Vantage platform, and warpage control
• VIEW Micro Metrology's high-throughput telecentric metrology across wafers and large panels

Learn more at imaps.org

Support the show

Episode Transcript

SPEAKER_11 (00:00):
This episode of the 3D Insights Podcast is sponsored
by IMAPs, the premier globalassociation for microelectronics
advanced packaging enthusiasts.
A membership in IMAPs helps yourcompany grow its advanced
packaging workforce throughprofessional education and
networking, advances your brand,and supports building
relationships.
IMAPS helps you learn, connect,and collaborate.

(00:20):
Learn more at imaps.org.
Hi there.
I'm Francoise von Traffe, andthis is the 3D Insights Podcast.

(00:41):
Hi everyone.
This week we are recording livefrom San Diego where I am
attending the IMAPSInternational Symposium.
Now, I realized last week thisis my 20th year at the event.
In fact, it was the first eventI ever went to in the advanced
packaging industry.
And I actually just announced myretirement from 3D Insights at

(01:03):
the end of the year.
So in this episode, I'll bespeaking with some of our 3D
Insights community members toget their perspectives of where
advanced packaging is going andtheir role in it.
And also what's up with theircompanies.
So my first guest has been heremany times before.
I'm speaking with Keith Feltonof Siemens Digital Industries
Software.
Welcome back, Keith.

SPEAKER_09 (01:24):
Thank you very much, François.
It's great to be here.
Great to see you again.
Probably for the last time,actually.

SPEAKER_11 (01:30):
Yeah, we'll see.
You'll see me.
Hopefully.
That's okay.

SPEAKER_09 (01:35):
As long as you're still the queen of 3D, just for
the different crowns.

SPEAKER_11 (01:37):
I'll always be the queen of 3D.
You can't take that away fromme.

SPEAKER_09 (01:41):
Wonderful.

SPEAKER_11 (01:42):
So you guys have some big news that you shared,
and it has to do with anotherone of our members, ASE.
You're collaborating on one ofyour products, Innovator 3D IC,
and it's driving 3D blocksworkflows for ASE's VIPAC
platform.
That is a lot of acronyms.
That's one of my favoritetopics.

SPEAKER_09 (02:02):
But we are acronym rich in this industry.

SPEAKER_11 (02:05):
So if you can maybe unscramble that Scrabble for us.

SPEAKER_09 (02:10):
So ASE have a um a platform of advanced integration
technologies, and they'vebranded it VIPAC.
Very similar to the way TSMChave 3D fabric.

SPEAKER_11 (02:21):
Okay.

SPEAKER_09 (02:21):
And they came to us and said, you know, we're trying
to drive adoption andproliferation of this
technology.
And we believe the way to do itis to provide our customers with
3D blocks syntax models of ourtechnology.
Okay, so 3D blocks is a languagedefinition format that lets you

(02:42):
describe how to put together anadvanced semiconductor package.

SPEAKER_11 (02:46):
So 3D blocks spelled 3D BLOX is a workflow.
Okay.

SPEAKER_09 (02:59):
So it's like a 3D language model that you can use
to give to someone else and theycan consume it electronically,
digitally.

SPEAKER_11 (03:09):
Okay.

SPEAKER_09 (03:09):
Okay.
It's written in YAML format ifyou know what YAML is.
It's a human readable languageformat.

SPEAKER_11 (03:16):
You know, EDA has always baffled me with all of
your I mean, you guys have yourown language.

SPEAKER_09 (03:21):
Well I don't think YAML was invented by EDA, but
it's Y A M L.
And obviously ML is somethinglike markup language.
Don't ask me what the rest ofthe YA means.
But young adults?
No possibly.
But it was created by TSMC.

SPEAKER_11 (03:38):
Okay.

SPEAKER_09 (03:38):
So they created it to better communicate
information with their customersinstead of doing it with um
design rule manuals, which arepaper printed documents.
Okay.
So they did it, they wanted tocreate something that was a
digital definition that theycould give someone, and they
wanted the industry to both readand write this data format.

unknown (04:00):
Okay.

SPEAKER_09 (04:01):
Because it allows interoperability.

SPEAKER_11 (04:03):
Okay.

SPEAKER_09 (04:03):
So, you know, they drove um a group of companies to
come together to help uh adoptthe format.
Siemens was part of it, everyEDA vendor was part of it.
Then when they got it to acertain level, they wanted to
get it turned into a bona fideindustry standard.
It was still proprietary at thetime.
So they worked with IEEE andit's now being adopted by IEEE.

(04:28):
So they've taken over.

SPEAKER_11 (04:29):
This is not an interface like UCIE.

SPEAKER_09 (04:32):
No, it's not.
No, it's not.
It's just I would think of it.
Remember Edith?
Oh.
I'm sorry to say that word.
Alright.
Do you know you know things likeJEDEC, IDX, IDF?
It's kind of like that.

SPEAKER_11 (04:48):
Right, I think that's right.

SPEAKER_09 (04:49):
But it's but it's really targeted for uh
semiconductor package assembly.

SPEAKER_11 (04:53):
Okay.

SPEAKER_09 (04:53):
So what happened was that ASE wanted to proliferate
the adoption of their VIPtechnology, and they believed
that doing so with 3D blocks,giving out 3D blocks definitions
of their advanced packages,would be a good way for people
to easily consume and adoptthem.
Because 3D blocks is read by allthe EDA vendors.

(05:16):
So they all read it.
So, for example, if I was tocreate 3D blocks in tool A, tool
Z from a completely differentvendor can read it because it's
a well-structured language, theycan read it and interpret it
correctly.

SPEAKER_11 (05:29):
I get it.

SPEAKER_09 (05:30):
Now the challenge is that you have to author 3D
blocks.
So there is a there is a writtenlanguage specification, and you
can type it by hand, it's YAML,which is human readable, but
that's not a very efficient wayof doing it.
So ASE in Kaoshung, um, led byDr.
CP Hung, came to our guys andsaid, Look, do you have anything

(05:55):
to help me digitally author 3Dblocks files?
And we said, Yes, we haveInnovator 3D IC, and you can
actually build um a digitalmodel of your uh package
assembly in that tool, and wecan export 3D blocks.
So we can read 3D blocks, we canauthor 3D blocks, we can edit 3D
blocks, and we can spit outexport 3D blocks.

SPEAKER_11 (06:18):
Okay, so the collaboration is between Siemens
and ASE on using your tool, theInnovator 3DIC, to enable them
to create the 3D blocksworkflows that they can then
give their customers theirproperty.
Okay, and 3D blocks is opensource.

SPEAKER_09 (06:36):
It's an open standard.

SPEAKER_11 (06:39):
No, oh no, no.
Right, okay.

SPEAKER_09 (06:41):
No, we're part of the language definition team, as
are all the other EDA vendors,but we sit on the committee and
we give them guidance.
But you know, now it's owned byIEEE, they have a very rigorous
and formal process to turn itinto a true global worldwide
standard.
Even though at the moment youyou could argue it's
semi-proprietary, still has alot of TSMC in it, right?

(07:03):
It's still very usable and itworks.
And so the team in Kaoshung atASE got to using Innovator 3D IC
to put together their various uhplatforms that they offer.
It's basically Focus, which is afan-out wafer level package,
Focus Bridge, which again isfocused, but this time with
embedded um silicon little minibridges to go between the die or

(07:26):
the chiplets, and then their twoand a half and three D middle
end of line uh solution, um,which is an interposer,
basically an interposer.
So for those three very advancedplatforms, they use our
innovator 3D IC software toprototype the 3D assembly so
they get it all correct, so it'smanufacturable, then they export

(07:48):
the 3D blocks files, then theycan give it to their customers.
And even if their customers donot use Siemens, the other EDA
tools can all read it.
So whether you're using acompetitive tool of ours, it
makes no difference.
And that's what I that's whatappealed to ASE because their
customers use all the EDAvendors.
They can't be one vendorspecific.

(08:10):
So they like the fact that it'svery neutral.

SPEAKER_11 (08:13):
Now, this is an ex not an exclusive license to ASE.
No, you could have the samearrangement down the road with a
ASEAN.
With everybody, yes.

SPEAKER_09 (08:22):
I mean basically, you know, ASE are one of our
OSAT Alliance members, as yousaid, so are AMCOR as well, as
well as other OSATs.
But no, any of any of them canapproach us for this.
ASE are just, in my opinion,they're very forward-looking.
And and typically OSATs are notforward-looking, right?
They're very conservative, um,but they're quite

(08:42):
forward-looking, and theybelieve this is the way they can
drive greater engagement withnew customers and prospects.
And we were very excited to workwith them.
You know, I've known CP um atASE in Kaoshung for many, many
years.
He's a very easy gentleman towork with, and um, they were
great.
So, you know, that's what weannounced last week.

(09:03):
And you know, ASC can deliverthat 3D blocks models today.
They're all fully ready to go.
So all customers have to do iscontact their ASC representative
and request a 3D blocks file forthe VIPAC technologies.

SPEAKER_11 (09:17):
Well, there you go.
Okay.
Before I let you go, I just wantto get a little input on where
you think advanced packaging isheaded because I feel like
advanced packaging is finallygetting the spotlight that it's
been deserving for so long.
So, what do you think is goingto be next now that people are
finally getting it?

SPEAKER_09 (09:35):
So, you know, a lot of people, especially in AI and
hyperscalers, edge node, andhigh you know, the what you
heard this morning, they arelooking at getting lower
pica-joule per bit.
That's what they really need.
It's all about power, right?
Because you look at some ofthese um data centers, they can
they you know, terawatts ofpower.

(09:56):
Right.
So 3D is the way to go, goingvertical, because then you get
the shortest metal pathway, andyou can lose much lower
voltages, so therefore your picajoule per bit is much lower.
The problem you have with that,one of the things the biggest
problems we see thermal andthermally induced stress on the
transistors, basically, becausethey're so close together now,

(10:19):
because you've stacked logicdevices, then the next is a
test.
Test is a real problem.

SPEAKER_11 (10:25):
This is always the same thing, it always comes down
to thermal and test.
Always, always, always.
I've been hearing this for ohgoing on 20 years.
So I'm looking forward to seeingwhat happens.
I know that the HIR roadmap isfocused on thermal.
I know that there's a wholeconference that's running
concurrent with this one onthermal.
So yeah, I think I think you'reright, and we'll be interested

(10:47):
to see where it goes.

SPEAKER_09 (10:48):
You know, traditionally you use um probe
cards with probe needles.
So the problem there is, youknow, the density of probe
needles can only go so far, andum, so people are struggling
with that.
Also, the probe pressure on diecan be quite significant, and
you're concerned about damagingthe die.
And then, secondly, is theamount of power you require to

(11:10):
test these devices.
We've heard from some customersthat they have to use a kilowatt
of power to test some 3Dchiplets, and that's excessive.
You can't do that, so you haveto end up using multiple pro
cards and testing the assemblyat multiple levels, which is
very expensive and timeconsuming.
So there are some realchallenges to overcome.

(11:32):
Design is always the easy part,it's fabrication test, and then
reliability.
That's the hard part.

SPEAKER_11 (11:38):
Somebody'll have to keep an eye on it for me.
Yes, okay, because I'm done.

SPEAKER_09 (11:43):
Don't rub it in.

SPEAKER_11 (11:45):
Thanks so much, it's always a pleasure.

SPEAKER_09 (11:47):
Pleasure to you again.
Thank you.

SPEAKER_11 (11:52):
So, my next guests today are from our premium
members, ACM Research, and wehave two repeat guests, Sally
Ann Henry and Jim Strauss.
But we have a very special newguest who has joined the company
probably six months ago, and hisname is Leo Archer.
Leo and I were business partnersa long time ago, but not since

(12:13):
I've been podcasting.
So this is a trip down memorylane.
Leo, welcome to the podcast.
Thank you.
And Sally Ann and Jim, ofcourse.
Thank you.
One of the themes this week hasbeen the silicon guys finally
catching on to advancedpackaging.
And Leo was previously on thefront end and now he's here

(12:35):
learning all about advancedpackaging.
So, Leo, tell me a little bitabout yourself and your new role
at ACM.

SPEAKER_06 (12:42):
So, uh spent almost 20 years in the front end,
everything from transistorthrough interconnect.
And uh for the longest time,that was the focus of
everything.
We're all driven by Moore's Law.
And when you and I got intobusiness a long time ago around
this, um, that was stillprobably pretty much the case,
but I certainly didn't foreseejust how the evolution of

(13:07):
packaging would take place.
And I was actually out of theindustry for almost a decade
doing other things, and I'm backin now.
And part of the reason I'm backin is because a company like
ACM, who's growing and has uhactually a broad range of
portfolios for a lot ofdifferent things, front-end and
back end, is also making hugeinroads into wafer lever

(13:29):
packaging.
So that of course sparked myinterest to see just what is
happening, and I think fromconferences conferences like
this and even some of the othersthat I was in there earlier this
summer, um, it's become veryapparent that to continue the
migration to next generationtechnologies, particularly
driven by AI, it's not going tobe on the transistor anymore.

(13:52):
We have reached the laws ofphysics limits, and unless we do
a dramatic change in designandor materials, that's not
going to change.
So it is now down to packaging.
And if we want to do a 4 to 5ximprovement annually, there's no
other way to do it withpackaging.
So I will state up front, get itout of the way, because I know
it was coming.

(14:12):
You were right, Francoise.

SPEAKER_11 (14:14):
Oh, but I said that before.
I love being right.

SPEAKER_06 (14:19):
That's my homage to your retirement.
No, it's you you foresaw itdefinitely when we talked about
3D Insights way back in the day.
Um, I think it was a very novelconcept and it may have been
ahead of its time at the time.
Absolutely.
But you stuck with it, and Ithink a huge change was the fact
that social media in generalbecame so ubiquitous, and so

(14:40):
everybody was involved in socialmedia, and that's just grown and
grown and grown.
And you've taken a concept thatwas so foreign to an extremely
conservative industry,particularly in the front end.
And you adopted it and made itmade it what it is today.
So I mean, kudos to you forthat.
There's no two ways about it.

SPEAKER_11 (14:56):
Thank you.
I mean, there's still thank you.
And I think it's really funnythough that Intel still puts up
the slide that says Moore's Law.
But now we include that secondparagraph that says maybe there
may be a time we're gonna haveto take these chips and stack
them to achieve the performancepower area and cost.
And I was like, huh.

SPEAKER_10 (15:13):
And I hadn't even read that, but I'm glad that's I
saw you I saw you taking apicture of that slide in the
presentation.
Right.

SPEAKER_11 (15:21):
So Intel Intel is sticking to their guns.
They did it all.
It was Moore's law.

SPEAKER_06 (15:25):
But it's not even funny, it's no longer now kind
of getting away from Moore'slaw.
And it was never really a law,it was an observation.

SPEAKER_11 (15:32):
It was an observation, yeah.

SPEAKER_06 (15:33):
So even the great Intel has to adapt and change
it.

SPEAKER_11 (15:38):
Right.
Well, and ACM in general startedout in the front-end processes,
but you guys have seen for awhile the opportunities and
especially in panel levelpackaging.
Sally, and last year was yourfirst I'm up symposium.
And they were talking aboutpanel level packaging, but that
conversation has reallyexploded, hasn't it?

SPEAKER_03 (16:00):
Exploded dramatically.
And I think what we're seeing isthere's no uniform uh standard
for panels.
I mean, you've got the 310 by310, 510 by 525, 600 by 600, 700
by 700.
So where is the actualstandardization?

(16:21):
There isn't any.
And for an equipmentmanufacturer, it makes it very
difficult to design for allthese different panel sizes.
It doesn't make it easy.

SPEAKER_11 (16:30):
Yeah, welcome to advanced packaging because there
are no standards, and peoplekeep asking for them.
And that's one of the thingsI've noticed from the silicon
guys coming into packaging, evenin last week's um executive
summit.
It was it'll be really helpfulfor us in the material side if
there was standards aroundadvanced packaging, and
everybody was kind of like, theythink there's a standard for

(16:51):
panel, it's 510 by 515 or 600 by600.

SPEAKER_02 (16:56):
Right.

SPEAKER_11 (16:57):
Um but then I just started hearing about the 310,
which is basically a square 300millimeter wafer, right?
So, yes, how is this proposing achallenge to develop panel
tools?

SPEAKER_10 (17:08):
So this is all about the economics, right?
So um Leo was talking aboutticks required for technology,
but the panel is all about justthe economics of it, right?
Because you can do so many moredevices on a panel than you can
on a 300 millimeter wafer.

(17:30):
As it turns out, 510 by 515 is agood size for being able to do
5x the number of chips.
600 by 600, maybe it's a littlebit better, but it doesn't
really buy you that much.
And 510 by 515 is the standardthat a lot of companies I think
Intel is going five by the same.

SPEAKER_11 (17:50):
That's right.
Yeah.
That's another thing I've heardrecently is the original
argument for panel levelpackaging was increased volume.
Now it's really the large sizeof the AI chips that it's a
requirement.

SPEAKER_10 (18:06):
Right.
Exactly.

SPEAKER_11 (18:07):
So because those chips got larger, you're losing
more space on the wafer.
And the other thing is that thatlittle burst towards panel in
the beginning that kind of likewent away for a while, was
because those volumes dropped,and they were never gonna need
that many devices that would beon the panel.
So now I'm hearing though, yousaid it, they said it last week,

(18:30):
that we need the volumes againat a smaller size.

SPEAKER_06 (18:33):
Two things stand out to me as kind of being an
outsider coming back in,particularly into this end.
First of all, the back end ofthe line is now the front end of
the line, because in a sense,the challenges that were faced
in the front end all that timeago, you're now facing in the
back end.
Standardization, a lot ofmaterials that have to be
treated, you know, we did 20years ago on the front end.

(18:55):
But the other thing that um whenyou look at AI by 2030 is gonna
be about 300 billion, threethereabouts, four years later is
expected to be a trillion dollarby itself.
So if you take the basic fivechipsets that are used for AI,
that's gonna require hugevolumes.
Every phone next generation isgonna have AI built into it, you

(19:17):
know, automotive, medical, allof these things.
So what went before is like yousaid, it may not have been as
critical to have as many.
All of a sudden, it's gonna bein everything and everywhere.
And so then the cost that Jimtalks about is gonna be a bigger
factor.

SPEAKER_11 (19:32):
And if they're gonna be going into our phones,
they're gonna have to shrinkagain.
Yes.
So we keep doing the same thingover and over.
Isn't that the definition ofinsanity?
Exactly.
So with the challenges of nothaving standards, how do you
design a tool to support aclient?

SPEAKER_10 (19:51):
The tools that we're building, uh we're purposely
building them to be able tosupport panels at 510, 515, as
well as 600 by 600.

SPEAKER_11 (20:01):
So do they have to be huge?

SPEAKER_10 (20:02):
They're very big tools.
They take up a lot of space in aclean room.

SPEAKER_11 (20:07):
And I remember when the 450 days, it was all the
front-end processes beingdeveloped on 450, but nobody was
thinking about the tools foradvanced packaging because they
were still thinking about tapeand reel and boats, and
everything was being deliveredon chips, and they weren't
realizing that wafer-levelpackaging requires the tools to
also be developed for thepackaging space.

SPEAKER_06 (20:26):
But think about if you know the average tool in
wafer level packaging 15 yearsago.
If they spent more than 150,000,200,000, they were they were
shocked.

SPEAKER_11 (20:36):
Right.

SPEAKER_06 (20:37):
Now they're spending what the front end costs.
And that's a paradigm shift ineverything.

SPEAKER_11 (20:43):
It really is fun to have watched all of this happen
over the last 20 years that I'vebeen in the industry.
Um and things don't move ever asfast as you think they might,
which is probably why I wasahead of things, because I was
like because I would talk topeople and they would say, Well,
this is gonna happen in a fewyears, and then it'd keep
getting pushed out.
And I was a newcomer to theindustry and I couldn't

(21:04):
understand why.
And I actually could see thatthe system level costs would be
reduced if more money was spentin the packaging side, but
everybody was still stuck on thepackaging as a cost adder,
packaging has to be cheap, andit took a really long time to
get people to realizedifferently.
And I think Subu Ayer actuallydid a really good job in the

(21:26):
beginning.
He is the first person I everheard to say talk about
disaggregating an SOC into itsfunctional parts and
reassembling them, and and thatis what is now chip technology
today.
So I want to give you guys achance to talk about some of the
new developments at ACM since welast talked.
We got a few more minutes left.

SPEAKER_03 (21:48):
Well, not necessarily related to um
obviously we have our panels, wehave copper plating, metal
copper plating for panels.
We also have uh bevel cleanedfor the panels, we have a new
product for wafers, which doesframe cleaning.
Um we also have um in thedefinitely on the front end is

(22:10):
uh track.
We have a track system with 12quoters and 12 developers that
you can connect up to an ASMLscanner.
Okay.
And we also have what we call acopper deplating system for
compound semiconductors, whichhas just recently been
introduced.

(22:30):
So we're basically doing like aopposite of ECP.
Okay.
Right.
Um I don't know how to bestdescribe it, but basically DCP.
Yeah.
Deplating copper.

SPEAKER_11 (22:44):
So I'm not really familiar in the compound space,
other than the fact that it'sI've always been told that
compound semiconductor is theoriginal 3D.
Um but why would we have todeplate copper?

SPEAKER_10 (22:57):
So I think you actually plate it, and then
there are spaces where you haveto remove copper.
So it's like a selective etch.
It's a selective etch.

SPEAKER_06 (23:08):
That's exactly what it is.

SPEAKER_11 (23:09):
And what's the purpose of it?

SPEAKER_06 (23:11):
It's just part of like uh almost like the mass
process where you're laying downa pattern.
Okay.
And you've laid down a bulkcopper film.
Okay.
And you've got a response.

SPEAKER_11 (23:18):
So this is a front-end process again, sort
of.
It's not a packaging process.
Well, it's more backends.
It's an in-between.
It's in between.
Okay.
Okay, and that is a newannouncement that you just
launched that tool.
Yes.
And then you also launchedanother wafer tool for the front
end, the track.
The track.
Okay.
Well, I enjoyed thisconversation.

(23:39):
Leo, it's so good to talk to youagain.
I mean, really, it's been toolong.
It is.
You know.
We should do this again.

SPEAKER_06 (23:46):
Life is life.

SPEAKER_11 (23:46):
Who knows?
Well, when you come back out ofretirement and uh, I am leaving,
you know, I'm leaving thatoption open because nothing's
worse than showing up atsomething when somebody says, I
thought you retired.

SPEAKER_06 (23:59):
Well, I think you know I think IMAP should hire
you to come back and do thesethings on a regular basis.

SPEAKER_11 (24:04):
Well, maybe they will, it'll make me an offer,
make it a good one.

SPEAKER_06 (24:08):
One you can't refuse.

SPEAKER_11 (24:09):
Yeah.
All right.
But in the meantime, we wish youhappy retirement.
Yes, absolutely.
Thank you.
I'm looking forward to it.
Thank you.
At least for a little while.
Thank you.

SPEAKER_06 (24:18):
For the next thing.

SPEAKER_11 (24:23):
My next guest is Phil Sundon from Shellback
Semiconductor, and they'reactually a fairly new company
with 3D insights anyway.
So I'd like to give him a chanceto tell us a little bit about
himself and the company andwhere they see advanced
packaging headed.
Welcome to the podcast.

SPEAKER_00 (24:39):
Thank you very much.
Thank you very much.
I'm happy to be here and happyto see you in person finally.
So it's been a good thing.

SPEAKER_11 (24:44):
I know it happens.
You know, we we only gettogether at these live events
most of the time.
And then it's a lot of zooming.

SPEAKER_00 (24:50):
Uh-huh.
And and it's interesting, ourpaths are always crossing in
some way, shape, or form.

SPEAKER_11 (24:54):
So it's a small industry.
That's one of the things I loveabout it.

SPEAKER_00 (24:57):
It is, it is.
And and I love about it too.
I mean, it's uh 30 plus yearsI've been in this industry.
So um so Philip Sundan, uh, I'mwith Shellback Semiconductor for
just over two years now, um,business development manager,
which uh essentially means Iwear a lot of hats.
I help the sales team uh engagewith customers on a technical

(25:17):
level.
Um I spend a lot of time inEurope with our sales director
there, Yannick Pilau.
Happy to be here at this uhconference, and I took a lot of
the sessions yesterday um to tryto get more up to speed on
things.
Um our involvement, or at leastmy involvement, has typically
been just with um UBM etch, umphotos of strip applications
back end of line.

(25:37):
Uh so yeah, I thought I'd takesome of the uh the learning
sessions yesterday.

SPEAKER_11 (25:42):
And yeah, that's one of the really good things about
IMAPs is their professionaldevelopment courses.
You know, I was at an IMAXexecutive summit last week.
I did a panel on strategies inthe future of advanced packaging
and and market trends and what'sdriving it.
And someone asked who was new tothe segment of the industry.
A lot of people who were therewere front-end people that are

(26:04):
now recognizing that advancedpackaging is important and
really the key to what's comingnext.
So someone asked, Well, how muchmore innovation will we need in
advanced packaging?
And and the panelists and I wekind of like looked at each
other, and I was like, Well, Ithink you're never gonna not
need innovation in advancedpackaging.
Someone else asked aboutstandardizing in advanced

(26:26):
packaging, and they've neverreally been able to standardize
advanced packaging.

SPEAKER_00 (26:30):
And and what is the standard panel size?
Uh we just heard that thismorning.

SPEAKER_11 (26:34):
Three.

SPEAKER_00 (26:35):
Wait, four.

SPEAKER_11 (26:36):
The semi-standard is 500 by 515 and then 600 by 600.
But they've heard 700 by 700 wehave.
Yeah, well, you know, you'll getthis.
Anytime you bring in a new sizeof anything, you need to
re-qualify the processes to makethat make sure they're gonna
work on the differentdimensions, right?
So you previously, beforeShellback, your history you said

(26:59):
is a photoresist strip and UBMetch, photoresist strip, metal
lift-off.
The removal processes, thecleaning processes and the
etching processes.
So, how have you seen thatshifted?
I mean, how that's always beenround processes, right?
How is panel impacting that?

SPEAKER_00 (27:17):
Interesting.
So uh several years ago I wasinvolved in some panel projects
with an ultrasonic company.
So um we had the fog nozzles andcurtain nozzles and things like
that.
So how do you spin a squaresubstrate and effectively coat
and cover all the way to theedges?

SPEAKER_11 (27:33):
Is it gonna be like a what is it like the slit coat?
It's more of a scan.
The slit coating?

SPEAKER_00 (27:37):
Yes.

SPEAKER_11 (27:38):
Okay.
So your you told me before wegot on that your specialty is
really around single waferprocessing.

SPEAKER_00 (27:45):
That is my history for 11 years before I came to
Shellback.
So Shellback is obviously batchspray, the legacy semi-tool
process.

SPEAKER_11 (27:53):
Well, I heard last week that according to Chat GPT,
single wafer processing is goingto be the way to go.
Tell me a little bit aboutShellback and what you guys
specialize in.

SPEAKER_00 (28:03):
That's a nice segue.
Thank you.
Um, Shellback, uh, we own the IPto the legacy semi-tool um spray
batch or batch spray umtechnology.
It's a full cassette of wafers,uh 200 millimeter, 100
millimeter.
It's a legacy process line thatis used throughout the world,
continues to be used throughoutthe world.
If somebody says I'm going topick up a gyroscope and start to

(28:25):
manufacture that, they'll lookat the history.
Okay, how do you manufacture agyroscope?
And their semi-tools name.
So it's just been aroundforever.
It's been the process tool ofrecord in so many of these.

SPEAKER_11 (28:35):
But this is a batch tool.

SPEAKER_00 (28:37):
It's a batch tool.
So yes, now we're talking aboutsingle wafer.
So there's advantages anddisadvantages, just like
everything.

SPEAKER_02 (28:43):
Okay.

SPEAKER_00 (28:44):
Um, spending 11 years in single wafer
processing.
I am a huge fan of Screen andTell as an equipment company.
They're the best in the world.
They're doing front-end of lineprocesses that nobody else can
do.
So, from a single waferstandpoint, there are there
certainly are advantages.
Um, single wafer processing isone wafer at a time.

(29:05):
So you can control theselectivity of the etch from
center to edge.
Um, you can certainly removethings uh evenly and cleanly,
where if you're doing a batchspray, you've got the wafers
sandwiched between each other.
So if uh different types ofmaterial is coming off, it might
get trapped.
So there's it just depends onthe process and what you're

(29:28):
trying to do.
But batch spray gives you thebenefit of spraying 25 wafers at
a time.
Right.
So the throughput is typicallymuch higher.
So on devices that are fairlysimple, um, if it's a metal
liftoff process, the metalcoming off is like glitter, I
always say.
Okay.
Then we can manage glitter veryeasily and very quickly.

(29:49):
So if it's a glitter removalprocess, great.
That's that lends itself to alot of people.

SPEAKER_11 (29:54):
I love that analogy.
I mean, I've been hearing aboutmetal liftoff for a long time.
And I kind of got it, but youjust made it crystal clear.
So thank you for that.

SPEAKER_00 (30:05):
Right?
Right.

SPEAKER_11 (30:06):
Who doesn't love glitter?

SPEAKER_00 (30:07):
Uh-huh.
Yes.
Uh uh.
Showgirl glitter.

SPEAKER_11 (30:11):
So do you have a single process tool?

SPEAKER_00 (30:14):
So uh certainly Semi-Tool did have a single
wafer process tool.
Um, it was a face-down process.
It was actually very successful.
And it was used in a lot ofapplications where, if you can
imagine, gravity lends itself touh a removal process.
So if that's beneficial, yes.
We currently don't make thattool, but we do own the IP to

(30:37):
it.
So we could manufacture thatagain, just depends on what's
happening.
So the business situation theway it is right now, we're
focusing on the batch spraytool.

SPEAKER_11 (30:46):
So did Shellback basically acquire all of the
technology of semitool?
Is it a rebrand of semi-tool?

SPEAKER_00 (30:53):
Uh not a rebrand necessarily.
So when Applied Materialspurchased semi-tool from Ray
Thompson, uh what they wereinterested in was the copper
plating technology.
Okay, yeah.
So they didn't really care aboutthe metal liftoff or the
cleaning or the etchingprocesses, they just wanted that
copper plating.
So they spun that off to a groupcalled OEM Group.

SPEAKER_11 (31:14):
Ah, okay.
You remember OEM groups?
Yes.
The history.

SPEAKER_00 (31:17):
Yes.

SPEAKER_11 (31:18):
Did we were we talking before about how small
this industry is?
It is, yes.

SPEAKER_00 (31:22):
Yes, yes.
So the owners of OEM Groupthought that they wanted to get
out of this industry.

SPEAKER_11 (31:27):
Yeah.

SPEAKER_00 (31:27):
You can never leave.
By the way.

SPEAKER_11 (31:30):
Yes, no, I understand that.

SPEAKER_00 (31:31):
They pretty much dissolved OEM group, sold off
things, but uh, there was apassion for the wet process that
remained.
And so they reacquired the wetprocess group that's in
Coopersburg, Pennsylvania, whichis where our headquarters is.

SPEAKER_11 (31:43):
I see, okay.

SPEAKER_00 (31:44):
So that is the original Retech facility, which
was owned by Semi-Tool.

SPEAKER_11 (31:49):
Okay.
And so all of the IP, as I lookon your website, I'm seeing
yeah, exclusive semi-toolproduct and IP owner.

SPEAKER_02 (31:58):
Yes.

SPEAKER_11 (31:58):
You've rebranded it, you've done different things to
the tools, improved them.
Um, one of the tools that yourecently announced, and that I
think you're advertising on 3DInsights right now, is something
called um the Tarrant.
It talks about hydrazone andfluorizone processes.

SPEAKER_00 (32:16):
Yes.

SPEAKER_11 (32:16):
Can you explain what those are?

SPEAKER_00 (32:18):
Yes, yes.
So in a nutshell, it is usingHydrazone as the accelerator for
a process.
So Hydrozone in itself is agreat oxidizer.
And what we are promoting iswhat Semitool developed, and in
fact, won the Best of the Westaward in 2009 at Semicon West

(32:40):
for this technology.
15 years ago.
Yes, yes.

SPEAKER_11 (32:42):
See, this is what I'm talking about.
I've been doing this for 15years.
I need a new story.

SPEAKER_00 (32:47):
So so Best of the West 15 years ago, but we had
this great product back then, orthis great material back then
called NMP.

SPEAKER_11 (32:55):
Okay.

SPEAKER_00 (32:56):
And everybody was using NMP for resist strip,
which is what hydrazone can beused for.
So they said, okay, this isgreat.
This is a neat little trick, butwe don't need to use this.
We've got NMP.
Well, now, 15 years later,people realize, oh, well, NMP
maybe isn't as great as wethought it was for the
environment.
So they are starting to comeback and look to us because

(33:16):
hydrazone is able to replacehazardous chemicals.
Essentially, what comes out ofthe tool at the end of the day
is hot water.

SPEAKER_11 (33:25):
Okay, so that is really interesting to me.
And also it just reminds meabout how sometimes it's about
timing.
Absolutely.
And you develop a technology andit doesn't take off, and it
generally has something to dowith we use whatever we have as
long as we can until we can'tuse it anymore.
And that was how a lot why ittook so long for 3D stacking to

(33:47):
happen.
A lot of these technologies arelike you said, no never goes
away.
Right.
But once there's a time for it,we pull it out and we
reintroduce it.
And the thing is, you have a lotof new people.

SPEAKER_00 (34:00):
Exactly.

SPEAKER_11 (34:01):
Because that's the other thing.
My other theme these days islike I've been telling the same
story, but it's a new audience.
It's always a new audience.
And right now, the audience isthe silicon guys that are
finally going, okay, maybe weshould take a look at advanced
packaging.

SPEAKER_00 (34:13):
Right.

SPEAKER_11 (34:14):
And so that made us legit.

SPEAKER_00 (34:15):
Right.
So right.
And it's, I mean, I spent 13years in Portland, Oregon, and I
became a little bit of a treehugger there.
I was already a tree hugger.
So I'm passionate about thishydrozome project.
Um, that's what I spend a lot ofmy time trying to convince
people that it is actually aviable process.
Um, process engineers are veryrisk averse, as we know.

(34:36):
Um, if it's working in the fab,nobody wants to upset the Apple
cart.

SPEAKER_02 (34:40):
Right.

SPEAKER_00 (34:40):
But yeah, so we've got the newer process engineers
that are coming in, and we'vealways relied on our semi-tool
legacy brand name.
The newer process engineersdon't know semi-tool, they've
never heard of semi-tool.
Right.
But let's talk about a greentechnology, Hydrazone.

SPEAKER_11 (34:54):
Exactly.

SPEAKER_00 (34:54):
That is of interest.

SPEAKER_11 (34:56):
So let me ask you this because we focused on
sustainability for a long time.
Um the industry is still workingon it, I believe.
The globally, everybody cares.
Right here in our littlecountry, suddenly sustainability
is a made-up story.

SPEAKER_02 (35:13):
Right.

SPEAKER_11 (35:13):
The risk of going down the geopolitical how does
that I mean, are are companiesbased in the US facing uh
elimination of thoserequirements, or are they
sticking with it because theyknow it's the right thing to do?

SPEAKER_00 (35:27):
Right, right.
And I mean, fundamentally youcan just use this as a um an
argument for cost reductionbecause solvents and formulated
solvents are extremelyexpensive.
DI water is expensive when youthink about the environmental
impact, but it is much lessexpensive than a solvent, a
formulated solvent.

(35:47):
I'm not bashing the solventpeople because we use solvents
in our spray solvent tool.
But if we can offset some ofthat, 10% of that, think of the
the impact of that.

SPEAKER_11 (35:57):
Right, exactly.

SPEAKER_00 (35:58):
So yeah, I'm not saying that hydrazone is going
to fit every single application,but certainly 10 to 20% of them
could benefit from thishydrazone technology.
So photoresist strip, that's aproven application.
Um, polymer removal, that'sanother proven application in
this industry here.
Um so it's it's definitely aviable technology that it's it's

(36:20):
worth talking about.

SPEAKER_11 (36:21):
Okay.
Well, that's all we have timefor today.

SPEAKER_00 (36:24):
Thank you for your time.

SPEAKER_11 (36:25):
Thank you.
Okay, so my next guests are fromDECA.
I'm speaking with Robin Davisand Tim Olson.
This is not their first rodeo.
They've been on the podcast anumber of times because DECA is
a premium member of 3D Insights.
Thank you both.
Thank you.

(36:46):
One of the things I've beenthinking a lot about lately is
how we've been telling the samestory about advanced packaging
for about 15 years, but theaudience keeps changing.
So they still have to learn whyadvanced packaging is so great.
But we finally find that thesilicon guys are starting to
realize that too.
And Tim, you and I were both atthe um IMAPS Executive Summit

(37:10):
last week in uh Silicon Valley,and I was actually having
trouble coming up with somethingto write about because it was
nothing was new to me.
But I realized that the audiencewas new.
So you've been in the industryfor a while.
Where do you see now thatSilicon Valley has finally
caught on, how do you see thatimpacting the future of advanced

(37:33):
packaging?

SPEAKER_04 (37:35):
That's a great question.
Much more involvement up anddown the supply chain.
And as you mentioned, a lot ofnew participants uh involved.
At that conference youreferenced, you and I uh were
able to listen to one of theleading private equity firms um
in semiconductors.
And what hasn't changed is evenwith all the advancements in

(37:55):
advanced packaging, as a privateequity investor, he said, I
really don't invest inpackaging, and I don't invest in
equipment for packaging.

SPEAKER_11 (38:03):
I thought he just said he didn't invest in
equipment.

SPEAKER_04 (38:06):
No, he was very negative.

SPEAKER_11 (38:08):
Very negative about advanced packaging.
So if you have an advancedpackaging startup, don't go to
the VCs for money.

SPEAKER_04 (38:13):
Exactly.
Equip money was incrediblynegative.

SPEAKER_11 (38:16):
No, he was incredibly negative about that.

SPEAKER_04 (38:18):
But he was also negative on packaging.
And he said it's changed a lotand it's much more attractive,
but it's still not attractive tothem.
And the reason is they'relooking for a thousand times
multiple when they invest in acompany.
And I think advanced packagingis a very exciting industry now,
and we're doing amazing thingswith the technology.
However, it is still many ways,it's in a support role to create

(38:40):
the system or the silicon.
It's acting as a supporting castmember to the silicon.
So advanced packaging is stillan enabling technology to make
our industry work.

SPEAKER_11 (38:52):
But it's very important.

SPEAKER_04 (38:54):
Of course.

SPEAKER_11 (38:54):
Now, you though are the CEO of what was an advanced
packaging startup that hassuccessfully transitioned out of
startup mode into a fullcompany, right?

SPEAKER_04 (39:11):
We are self-sustaining, but I feel
we're still a teenager.
We're sev 16 years old.
And yes, my damn teenager.
Exactly.
So I was at Amcore before Istarted DECA, and the famous uh
founder and chairman of Amcore,Jim Kim, came and had dinner
with me when I resigned to tellhim I was starting a DECA.
He said, Tim, do you realizethat 98% of all startups fail in

(39:34):
the first 10 years?
And at dinner I got all animatedand said, Wow, 2% make it?
And he knew that I'd be aproblem.
So I we're I'm really proud tosay DECA is part of that 2% that
made it.

SPEAKER_11 (39:47):
Well, I remember when DECA launched because that
was one of the first things Idid with 3D Insights was to be
part of that press corps thatwent to the launch party.
And even though at the time Iwas focused mostly on 3D
integration, there was still alot of exciting things and also
the potential for whatever DECAwas putting out.

(40:09):
At the time, you just had yourchipscale package, but you had
the plans for M series and youhad the plans for um, I don't
know if you had the plans foradaptive patterning already.
We didn't.
But oh, but you were very quiet.
We learned all about making winethat night.
Didn't learn a whole lot aboutDECA, but so I'm super excited

(40:30):
to have been part of thatjourney with you.

SPEAKER_04 (40:33):
Very excited to have you along the whole time.

SPEAKER_11 (40:35):
So the journey continues.
And the last time we talked atECTC, you had big news with IBM,
and then you just recently madeanother announcement.
And Robin, you were instrumentalin that, I believe.

SPEAKER_07 (40:48):
Yeah.
Um, the whole team was reallyinvolved, but uh, we released a
press release with SST a coupleweeks ago now, announcing an NVM
chiplet package that we'reputting together with them, and
that is enabling them to pullout what they've traditionally

(41:09):
provided as IP that theircustomers can implement in their
chips into a separate chipletpackage, take it out of that
silicon and allow that siliconto scale to finer nodes while
keeping the memory separate.
We'll be helping them implementthat in our technology and prove

(41:31):
out the solution so that theircustomers don't have to redesign
the package from the ground upand they can just almost plug
and play their device and do it.

SPEAKER_11 (41:42):
So let's back up a little bit for people who don't
know, and that would be me.
Who is SST?
What do they do?

SPEAKER_07 (41:49):
Uh so SST stands for Silicon Storage Technology, and
uh they're a subsidiary ofMicrochip.
They provide memory IP.

SPEAKER_11 (42:00):
So are they a design house then?

SPEAKER_07 (42:02):
Yes.
They're IP in the same way thatum I mean DECA's IP as well.
Yes, but we're physicalimplementation.
Okay.
And they develop the siliconstructures that are going to be
implemented.

SPEAKER_11 (42:14):
Okay.
And then they use your IP to dothe packaging.
Yes.
Very cool.
They're part of Microchip.

SPEAKER_07 (42:23):
They're a subsidiary of Microchip.

SPEAKER_04 (42:25):
They were a startup many years ago in Silicon Valley
and they did a great job ofcarving out a niche in the
market to be the leading non-volNVM is non-volatile memory, but
permanent memory, independent ofthe power being there or not.
So they are the number onecompany in the world that makes
this type of memory technology.
And it's important because whenyou touch on on a semiconductor,

(42:46):
uh the on button, the powerturns on and it has to go read
its instructions from somewhereof what to do.
That's where it comes from, isfrom that NVM memory.

SPEAKER_02 (42:54):
Okay.

SPEAKER_04 (42:55):
So they created this soft IP that went inside other
people's chips and um for thelast couple decades.
And then Microchip acquired themuh several years ago, and
they're but they're asubsidiary, but they operate
independently.
So they still have their own setof customers, in many cases
competitors of microchip,actually, other microcontroller

(43:15):
companies and other analysts.

SPEAKER_11 (43:16):
Wrapping my head around this, the difference
between volatile memory andnon-volatile memory is that one
is needs power.

SPEAKER_04 (43:24):
Yeah, DRAM DRAM is classic volatile memory.

SPEAKER_11 (43:27):
Right.

SPEAKER_04 (43:27):
And the beauty of DRAM is it's super fast and it's
relatively low power, so that'swhy DRAM's big in AI and all
those things.
But um and in the case of DRAM,you constantly not only you have
to have the power on, youconstantly have to refresh it.

SPEAKER_11 (43:41):
So it's when you save it, so let's say you have a
Word document and you're writingaway, did power goes out and you
lose that whole thing becauseyou didn't save it, because it
was just the DRAM.

SPEAKER_04 (43:53):
That would be a problem.

SPEAKER_11 (43:53):
But if you saved it, then you've transferred all of
that data information over tothe nonvolatile memory.

SPEAKER_04 (44:00):
Exactly.
And in that case, you'reprobably transferring it to a
solid state disk drive in yourlaptop.
And that is a different type ofnonvolatile memory called NAND
flash, which you've probablyheard of.
And NORFlash, those so there'sdifferent types of NVM.
Um but uh the stuff in your harddrive is built by other
companies, uh, so that's whereyour Word document would go,

(44:20):
would be on that othernon-volatile memory.

SPEAKER_07 (44:23):
SST's NVM they refer to as EFlash for embedded flash,
embedded within the chip itself.

SPEAKER_11 (44:31):
Okay.

SPEAKER_07 (44:32):
Or embedded within the chiplet package now.

SPEAKER_11 (44:35):
Right.
Okay.
You've got the chiplet package,but only part of that is the
memory.
There's also logic correct chipsin the package.

SPEAKER_07 (44:43):
Historically SST's IP was being implemented in the
chip.

SPEAKER_11 (44:48):
Right, so it was an SOC.

SPEAKER_07 (44:50):
Yeah.
And now it's going to be splitout from that.
Okay.
Memory is its own chiplet.

SPEAKER_11 (44:55):
So that is a good example for anybody who doesn't
understand the whole concept ofdisaggregation of an SOC.
An SOC is a system on a chip,and you have the logic and the
memory, and maybe the RF device,and I don't know, anything
digital, right?

(45:17):
Is you don't can you haveanalog?
Yeah, you can have analog and soclearly.
All right.
So then it's a function ofdesign, right?
You design it separately but towork together.
And then that is packaged, sothat's where DECA comes in.

SPEAKER_04 (45:34):
Yeah, absolutely.
It's really interconnecttechnology and it becomes a
package, but it's hooking thatNVM chiplet that it needs
hundreds and hundreds of wires,and it's a tiny chiplet.
It might be probably less thantwo millimeters in dimension.
And it has a couple thousandwires potentially back to the
chip because you've got a lot ofinterconnect to the brains of
the chip.

(45:55):
In a lot of cases, that's amicrocontroller.
The most typical SOC that it'sbeen used with is MCUs.
So all the leading MCU companiespretty much use SST.
And then a lot of other analogDSP companies that also use uh
use that.

SPEAKER_11 (46:10):
So this is a big win for you guys.
Big big win.
Congratulations.

SPEAKER_04 (46:15):
Thank you, Russ.

SPEAKER_11 (46:16):
I think we're gonna have to call it a day.
I hope I get to talk to you guysagain.

SPEAKER_04 (46:20):
I hope so too.
We're gonna miss you.
You're retiring, my goodness.
What's the industry?

SPEAKER_11 (46:25):
But does anybody ever retire from this industry?
I'm retiring from 3D Insights.
Let's see what happens next.
I'm gonna take a little break.
Write a novel.

SPEAKER_04 (46:39):
Okay.

SPEAKER_07 (46:40):
I love that for you.

SPEAKER_04 (46:42):
Perfect.
All right, you guys.

SPEAKER_11 (46:44):
Good to talk to you.
Thanks.

SPEAKER_04 (46:45):
Thank you.

SPEAKER_11 (46:50):
So I'm speaking with Liam Thorpe from Norton
Corporation, and Norton is oneof our premium members.
Welcome to the podcast, Liam.

SPEAKER_05 (46:58):
Hi, thank you for having me.

SPEAKER_11 (46:59):
So I was excited to talk to you before we started
the conversation and learn thatyou are brand new to the
advanced packaging industry.
So that's our favorite people totalk to to get your impressions
so far.

SPEAKER_05 (47:09):
I'm excited to give impressions, frankly, very
excited to be in advancedpackaging.
It's really there's just so muchgoing on.
It's been it's been really afantastic kind of place to
start.

SPEAKER_11 (47:21):
So where did you come from?

SPEAKER_05 (47:24):
Yeah, so I I previously worked in uh for
about 18 months out of collegein the automotive space uh doing
design and release engineeringfor uh electric vehicles.

SPEAKER_11 (47:36):
Oh, okay.
So where were you based?

SPEAKER_05 (47:38):
I was based out of Irvine, California.

SPEAKER_11 (47:40):
Okay, so not you weren't working for Tesla, were
you?

SPEAKER_05 (47:42):
I was not working for Tesla, no.
No.

SPEAKER_11 (47:46):
So glad to hear that.
So what led you to Nordsen?

SPEAKER_05 (47:51):
So I um I was looking for uh new opportunity.
Um unfortunately, electricvehicles is a little volatile.

SPEAKER_11 (47:58):
Um depends on who you are.

SPEAKER_05 (48:03):
So they uh they handed the uh our engineering
department, the Sia Never, uhkicked us out, and I uh I ended
up looking for something newkind of in that space.
I really had enjoyed with thedesigner release engineers a lot
of work in kind of coordinatingacross systems with customer
input and you know basicallymarrying hardware, software, and

(48:27):
and user experience.

SPEAKER_01 (48:28):
Yeah.

SPEAKER_05 (48:29):
Um especially in in the uh exterior space, which is
what I was working on.
So I really was looking forsomething in that area, and
that's what what pointed me toNordson, and uh that's where I
ended up.

SPEAKER_11 (48:41):
And what did you study in university?

SPEAKER_05 (48:43):
I'm a mechanical engineer.

SPEAKER_11 (48:44):
Mechanical engineer, okay, because we know this
industry, you can be amechanical engineer, you can be
a materials engineer, you can bea chemical engineer.
The world is your oyster whenyou're an engineer.
So advanced packaging is anexciting space, I've always
thought.
One of the big topics here thisweek, and also basically in

(49:05):
general, is panel levelpackaging.
And I understand that you'reworking in the panel level
packaging space at Nordson.

SPEAKER_05 (49:12):
Yes, yes, we are.

SPEAKER_11 (49:14):
So how's it going?

SPEAKER_05 (49:15):
It is very exciting.
Um we are working uh reallydiligently right now, taking a
lot of customer input um anddeveloping kind of systems to
leverage some of our existingexcellence um in and around our
vantage platform, especially,um, and work that forwards to
provide a competitive edge forour customers.
We're still kind of in thedevelopment and deployment

(49:38):
stage, um, but really lookingforward to being able to present
that to all of our customers uhas a as an opportunity.

SPEAKER_11 (49:46):
So, as an equipment manufacturer, because that's
what you do, right?
You're doing dispense equipmentand also plasma equipment.
One of the questions out thereis about standards around panel
size, and that we really haven'tstandardized.
Their most focuses on 500 by 515or 600 by 600, but now I'm also
hearing about 310 by 310, andI'm hearing about 700.

(50:12):
What is challenging for anequipment supplier in having to
accommodate the potential ofdifferent size substrates?

SPEAKER_05 (50:22):
It's been really interesting, especially in the
space of, you know, withmanufacturing hardware, it's
especially hard because youcannot change the size of your
hardware once you deploy it tothe field.
And we really want to besensitive of customers trying to
optimize their manufacturingsizes and processes, and that
means making a machine with thelargest possible area with the

(50:44):
smallest possible footprint.
So that has been a reallyinteresting sort of uh challenge
for us to make sure we'redelivering in those larger panel
level sizes that have been uhkind of on the cutting edge and
also not, you know, delivering abehemoth of a machine that
couldn't be used for uh aproduction space.

SPEAKER_11 (51:02):
Aaron Ross Powell Right.
So are you focused on one ofthose sizes?

SPEAKER_05 (51:06):
We are not focused on any size in particular, but
rather uh the ability to spreaduh across a 310 by 310 all the
way up to 600 by 600.
We're working in kind of all ofthose.

SPEAKER_11 (51:18):
So nobody's really touching the 700s at this point.

SPEAKER_05 (51:21):
Not that I've seen, no.

SPEAKER_11 (51:23):
What are the challenges in developing
materials for panelapplications?

SPEAKER_05 (51:28):
I think especially with materials, there are so
many different ways we canapproach dispensing.
Um and to that end, choosingmaterials that especially meet
the customer application isobviously very important, but
then making sure the equipmentis tuned to that material or
that application has also beenuh I'm not gonna say only a
challenge for us.

(51:49):
It's also something we've reallyexcelled at and it's been quite
fantastic.
But with all of the constantupdates and improvements and new
materials, we see all kinds ofexciting new applications, and
some, you know, we're preparedto meet, and we're very, you
know, excited to be prepared tomeet that.
And some, you know, we'relooking at like, ooh, maybe that
is uh a new direction we need tobe prepared to go in for for

(52:09):
panel level or for otherapplications, especially.

SPEAKER_11 (52:13):
What about different substrate materials like um
dealing with dispense on theglass core substrate?
Is that something that you areinvestigating?

SPEAKER_05 (52:23):
Yes, yes.
Uh lots of different uhapproaches to how we want to
handle different panels uh andwafers, frankly.
You know, we're we're trying tostay flexible and and in tune
with our our different customersto provide with the hardware
they need and the fixation theyneed to operate, you know,
whatever they want to do to pushthe cutting edge.

(52:44):
We want to be there to supportthem.

SPEAKER_11 (52:45):
Okay.
Okay, so you want to tell usabout some of your new products
in the market?

SPEAKER_05 (52:50):
Yeah, so I've been very excited over this uh
summer.
We've uh recently released theIJ, uh the IntelliJet 1.1.
Um it's got some excitingimprovements to um our
maintenance process so you canrun it longer with less uh
downtime for maintenance.
We've been very pleased withthat.
And we're doing a lot of work inthe vantage space, in making

(53:11):
improvements to our flagship andand presenting it in different
uh applications and ways toreally provide that flagship top
of the line experience to ourcustomers, especially where
they're looking to push theenvelope in packaging and
underfill.

SPEAKER_11 (53:25):
Okay, and for people who might not be familiar with
your tools um by name, can youjust explain what the Vantage
does and what the IJ does?

SPEAKER_05 (53:34):
Yes, yes.
Uh our IntelliJet is apiezo-driven uh jetting valve,
and the Vantage platform is ouractual dispensing machine.

SPEAKER_11 (53:44):
Okay, so the IntelliJet works on the Vantage.
Yes.
And it runs on the Vantageplatform.

SPEAKER_05 (53:50):
Yes, as many of our valves do.

SPEAKER_11 (53:52):
And the um and you're focused on the panel
space for this?

SPEAKER_05 (53:56):
Yes.
Okay.

SPEAKER_11 (53:57):
But you said you still also do obviously you're
you still do wafers.
Yes.

SPEAKER_05 (54:01):
Absolutely, we still do wafers, and we still have
plenty of work going on in thewafers space.

SPEAKER_11 (54:06):
Oh, I'm sure, because you know it takes a long
time.

unknown (54:09):
Yeah.

SPEAKER_05 (54:09):
Yeah, especially in uh in warpage control.
And that's been something thatwe've been able to sort of take
lessons learned in in panellevel packaging and in wafer
level packaging, the warpage isvery important.
So we've been doing a lot ofwork in how do we minimize the
warpage for our customers sothey can work with these really
advanced substrates at differentmaterial or different
temperatures.

SPEAKER_11 (54:29):
And I'm sure depending on the size, it's
gonna impact the potential forwarpage.
Trevor Burrus, Jr.

SPEAKER_05 (54:35):
Exactly.
That's a big struggle in thepanel space, is you know, when
we're talking 600 millimeter by600 millimeter, that is a very
large panel, and a very small,you know, warpage towards the
center is a very large concernas we get towards the edges.

SPEAKER_11 (54:49):
Aaron Powell For me, this harkens back to the
transition between 200millimeter and 300 millimeter
wafers, where there wereprocesses that were developed
and qualified and functioning onthe 200 millimeter wafer, such
as temporary bond debond,because they were doing it in
MEMS and they were doing it incompound semi.
And then when they startedthinking about doing it for

(55:10):
stacking wafers and using TSVs,suddenly that D bond process
wasn't working at 300millimeter, which worked fine.
So I that's always stuck in mymind as something you need to
think about when you're dealingwith process across different
size substrates and differentusing different materials.

SPEAKER_05 (55:28):
Yeah, exactly.

SPEAKER_11 (55:29):
I'm sure it's something you're very aware of.

SPEAKER_05 (55:31):
Yes.

SPEAKER_11 (55:32):
Okay.
Well, I'm so happy to have metyou and thank you for joining me
today.
And um, you know, thanks forbeing part of 3D Insights.

SPEAKER_05 (55:42):
Thank you.

SPEAKER_11 (55:43):
Take care.
My next guests are Al Emmons andKen Peters from View
Micrometrology.
They are our newest members andthey are premium members.
Welcome to the podcast, guys.

SPEAKER_01 (55:57):
Thank you, Francois.

SPEAKER_11 (55:58):
So, because you're new and we've haven't really had
many conversations about what itis that you do and what brought
you to the advanced packagingindustry.
Let's start by if each of youjust give us a little bit of
your background and how youended up at Vue.

SPEAKER_01 (56:12):
Sure.
I started with Quality VisionInternational, the Parent
Corporation in 2021.
And uh I came from apharmaceutical wholesale
background.
Um so this was my first stint inoptical metrology.
So I learned from the ground upwith a division called OGP, they
make a product called SmartScopes.

(56:33):
I started in that division, andlast October, it's been about a
year now, Ken recruited me intothe Vue division.

SPEAKER_11 (56:40):
Okay.
So you they're using metrologyin more industrial applications
in the machining industry?
Yes.
And so how do you find the shiftinto the semiconductor space,
the microelectronics space?

SPEAKER_01 (56:53):
I am amazed at how micro these parts are and these
features are.
You know, compared to machinedparts, for instance, uh razor
blades, compared tosemiconductor chips.
I'm I'm just amazed.

SPEAKER_11 (57:07):
It's cool, isn't it?

SPEAKER_01 (57:08):
It just blows my mind.
Every day I learn something new.

SPEAKER_11 (57:11):
Right, yeah.
And Ken, how about you?

SPEAKER_08 (57:14):
Number of years in the uh dimensional metrology,
that's where OGP comes from.
Took a 15-year break, actuallystarted in 1989 as a young
engineer.

SPEAKER_11 (57:24):
Where did you go to school?

SPEAKER_08 (57:25):
Geneseum.
It's in New York State, it's astate school.

SPEAKER_11 (57:28):
Okay.

SPEAKER_08 (57:29):
It's one of the best state schools, actually.
Right.
So and I came back in 2000.
I did 15 years of embeddedimaging consulting.
And in 2010, came back to workfor the owners.
Okay.
So now I'm kind of a utilityexecutive with a strong
technical background, and I'vehad different roles within the

organization (57:50):
CIO, you know, Go Help Service, those sorts of
things.
And then when the CHIPSAT hit,they wanted to grow Vue in the
States, in North America, andEurope.
And what do we do to replicatewhat we already have in Asia?
You know, 75-80% of our machineswere sold into Asia directly
into the semiconductor industry.

(58:11):
We want to replicate that intothe states.
And that's where we started uhVue Micrometrology Center of
Excellence out of Rochester,which is where all the machines
are manufactured.

SPEAKER_11 (58:22):
So how long has Vue itself been established?

SPEAKER_08 (58:25):
Established maybe about 40 years ago.
Okay.
Quality Vision purchased them inlate 1990s.
And so we incorporated thattech.
They were from Simi Valley.
We moved that to Tempe, Arizona.

SPEAKER_02 (58:38):
Okay.

SPEAKER_08 (58:38):
And over the years incorporated their technology
and our technology together.

SPEAKER_11 (58:43):
So you've been in packaging for a while.
We have.
But this is your first time atIMAPS.

SPEAKER_01 (58:48):
Yes.
First time exhibiting.

SPEAKER_11 (58:51):
Okay.
So you've been here as anattendee or presenter before?

SPEAKER_01 (58:54):
I came to DPC in the U.S.

SPEAKER_11 (58:57):
So what made you decide that this was the right
market for you?

SPEAKER_01 (59:02):
Just that it's focused on packaging.
The people that attended DPC areextremely knowledgeable people.
They're involved in theprocesses that we can do
metrology in.
And I'm just impressed with theIMAPS organization overall.
It's a small community, verytight, very professional.
So coming from thepharmaceutical industry, I'd

(59:25):
like to be in small communitieslike that.

SPEAKER_11 (59:28):
What did you think of the show this week?

SPEAKER_01 (59:30):
I've had a great time in the booth.
We've met so many people, had somany quality conversations.
One of my goals coming here as amarketer was to say, do we
actually have a market?
And the answer is yes.

SPEAKER_11 (59:43):
Absolutely.

SPEAKER_01 (59:44):
People are interested in our products, and
so I'm very pleased.
It's been a great show for us.

SPEAKER_11 (59:49):
So you actually had a tool in the booth.
What were you showcasing topeople?

SPEAKER_01 (59:54):
That's uh one of our small benchmark systems.
It's running just a very simpledemo showing.
Multiple parts that we can runmetrology applications on.
It's not really representativeof our mainstream products, it's
more of a show system.

SPEAKER_11 (01:00:08):
Okay.

SPEAKER_01 (01:00:08):
It displays our optics.
The power of our products are inthe optics.
It's the same optics in all thesystems that we make.
And uh the optics have a numberof different features.
For instance, we do high-speedstrobing, so our systems offer
really high throughput as wellas high accuracy.
Our optics are fully telecentricin the full field of view, so

(01:00:31):
there's literally no distortion.
And you can snap images andcapture many more features than
other systems we're competingagainst.

SPEAKER_11 (01:00:39):
So are they intended for the RD side of things or
manufacturing?
Volume manufacturing?

SPEAKER_01 (01:00:44):
They can certainly be used in RD.
However, our core strength isthroughput, speed, okay, and
accuracy combined.
So our systems are running inAsia 24-7 in many companies.

SPEAKER_11 (01:00:57):
For the metrology, what are you measuring?
Are you measuring materials?
Are you measuring whole devices?

SPEAKER_08 (01:01:04):
Well to simplify mostly uh dyes, studs and dies,
and chiplets.
So we want to know the locationand the orientation of dies,
whether that's on a 200millimeter, 300 millimeter, or
uh panel, large panel, 600 or700 millimeter panels.
And because of the makeup of ourmachines, you can measure those
dies accurately over 800millimeters anywhere on the

(01:01:27):
stage and of course anywherewithin the field of view.

SPEAKER_11 (01:01:30):
So the stage can support different wafer sizes
and different panel sizes.
Awesome.

SPEAKER_08 (01:01:37):
Here we'd like to see what we've done a lot of
wirebonding, one of kind ofthings over the years, and with
the fast advancing of packagingand whatnot were one of our
tasks.
What other areas could weexploit with our existing
technology that we really didn'tthink about before, coming from
dimensional metrology andmachining.

(01:01:59):
That's where the heart of ourbusiness.
We're in our 80th year ofbusiness.

SPEAKER_11 (01:02:03):
Wow.
That's older than thesemiconductor industry.

SPEAKER_01 (01:02:06):
Oh, yeah.
Started in the machiningindustry in 1945.

SPEAKER_11 (01:02:11):
So let me ask you you mentioned that part of your
motivation of getting into thisindustry was to build up your
business here in the UnitedStates because of CHIPS Act
funding.

SPEAKER_08 (01:02:21):
It was more or less to exploit that, to see what we
could do to expand the overallbusiness.
We represent 10% of the businessfor Quality Vision
International.
That's the manufacturer officesin a 7.2 acres in Rochester, New
York, all hand built machines.
90% of our machines arevertically integrated, all the
way from making the optics, allthe machining, all the

(01:02:44):
electronics, the mathematics,and so on and so forth.

SPEAKER_11 (01:02:47):
So you're seeing the recipients of Chips Act funding
as potential customers?

SPEAKER_08 (01:02:51):
Absolutely.

SPEAKER_11 (01:02:51):
Right.
Have the recent events that havechanged whether or not Chipsact
funding is going to be dispersedimpacted your business at all?

SPEAKER_08 (01:03:00):
No.
No.
These are long things.
You're going to put a factoryand it's not done over a
weekend.
So you're going to make amulti-year commitment for those
sorts of things.
You may pause it, you may dothose sorts of things, but
interesting, it's the samecompanies in Asia that are
building out into Europe, intoNorth America.
Yes.

SPEAKER_11 (01:03:20):
Well, it's a global industry, right?
Of course.
And that's, I think, one of thethings that I've always found
cool about the semiconductorindustry is that it requires
collaboration across borders andacross, you know, nations.
And that one of my concerns ishow in the last few years there
has been this shift, everybodywants to build out their own

(01:03:40):
ecosystem to try to get marketshare as a country as opposed to
a company, right?
If China builds out itsecosystem and the US builds out
its ecosystem and Europe buildsout eco its ecosystem, is there
going to be so much competitionthat one region is going to win
out over the other?

SPEAKER_08 (01:03:56):
I don't see it that way.
You know, the fabric of a lot ofthese international companies,
they're spread out all over theplace.
A company isn't gonna unrooteverything they got in China and
Taiwan, or they're gonna plop itin Arizona and say, okay, we're
gonna all live here.
That's not gonna happen.
We're gonna have a little bithere, a little bit here, a
little bit here.
It's much cheaper to localizesome of that manufacturing,

(01:04:19):
anyways.
Right.
Right.

SPEAKER_02 (01:04:21):
Right.

SPEAKER_08 (01:04:21):
And of course, most of the workers, right?
Most of that stuff's done inTaiwan.
Why?
Because you have to find theworkers that are capable of even
working within those industriesfor these kinds of uh in the
semiconductor, OSAT orotherwise.
So I don't really share the viewthat everything's gonna go back
as if the world doesn't existanymore.
It's a very, very small world.

(01:04:43):
In the last hour we talked topeople from, you know, Korea,
right here in IMAPS.

SPEAKER_11 (01:04:48):
Right.

SPEAKER_08 (01:04:48):
Right, right.
Korea, Japanese.

SPEAKER_11 (01:04:49):
Oh yeah, well that's what I mean.
This is a global industry, andit and it does best when it
stays global.

SPEAKER_08 (01:04:54):
We do business in China.

SPEAKER_11 (01:04:55):
Right.

SPEAKER_08 (01:04:56):
Right?
We're doing business inMalaysia, yeah, Vietnam.

SPEAKER_01 (01:05:00):
Some of our customers in the US are from
Asia.

SPEAKER_02 (01:05:04):
Right.

SPEAKER_01 (01:05:04):
So they're still operating in Asia and now
they're establishing operationsover here, so it's a cross full
global project.
So I completely agree.

SPEAKER_11 (01:05:16):
So busy week this week.
Next week we're all at SemiconWest.
You'll be there.
We'll see you there.
And uh thanks for joining metoday, guys.
Thank you.

SPEAKER_01 (01:05:25):
Thank you for having us watching.

SPEAKER_11 (01:05:30):
That wraps up this year's coverage of the IMAP
Symposium, but don't go away.
Next week we'll be back withcoverage from Semicon West and a
discussion on a nationalworkforce development initiative
and a recap of semisustainability content from the
show.
There's lots more to come, sotune in next time to the 3D
Insights podcast.

(01:05:50):
The 3D Insights Podcast is aproduction of 3D Insights LLC.