An Interview With:
Steve Silverman
Bartlett Bay Consulting
Barbara Kanegsberg
BFK Solutions LLC
Steve Silverman is a consultant who works with semiconductor and pharmaceutical
manufacturers to achieve contamination control in cleanrooms and mini-environments.
He provides practical approaches for avoiding contamination problems as
well as approaches to achieving multi-use cleanroom facilities. Steve
also provides
thought-provoking comments concerning contamination specifications, the
future of miniaturization, and the perennial dichotomy between training
and practice.
Those involved in a range of manufacturing applications will find practical
utility in his comments.
How did you become involved in contamination control? (BFK)
Silverman: In the late 1980’s, I became involved in a project dealing
with electron beam technology. The process was particle-sensitive.
As the personnel were not aware of the contamination issues, they had
no idea how
to solve the problem. I made it my mission to learn more about contamination
control; I found it to be an interesting field, and made it my new
career.
My consultancy grew out of that decision: I became involved in the development
of mini-environments. Bartlett Bay Consulting encompasses three areas:
design, problem solving, and contamination metrics. My favorite projects
involve
problems that we for-see, namely, fixing and addressing problems
before they happen. Of course, sometimes I am also called in to fix
problems
after the
fact. While we have some projects in the pharmaceutical industry,
our primary focus is contamination control of tools with mini-environments
in the field
of semiconductor fabrication.
What would you say is the most common, avoidable contamination control
problem? (BFK)
Silverman: The biggest avoidable problem that I see is that contamination
control is the last thing the engineer considers when designing
a semiconductor fabrication tool. They design themselves into a
box,
and cannot achieve
the required particulate level. Then, they have to call in my company,
Bartlett
Bay Consulting.
Before 1990, everyone depended on the cleanroom itself to keep
things clean, and the industry relied on the Cleanroom/Fab
Manager for that
task. In
the late 80s, when it became apparent that cleanrooms could
not keep pace with
the tough contamination levels needed, mini-environments became
necessary to meet contamination control specifications. That
changed the focus,
and the emphasis in responsibility for contamination control
of wafer tools
shifted from the Cleanroom Manager to the OEMs. As current
needs for contamination control is now at the ISO-1 level (a difficult
level
to achieve), OEM’s
are asking Bartlett Bay Consulting to provide solutions to achieving
contamination control by designing and testing such mini-environments.
The desirable approach would be to have the tool designer ask,
right at the beginning, “How can I keep this tool clean?” Instead,
the primary emphasis is on the semiconductor process, and contamination
control becomes
an afterthought. Consequently, the resulting contamination
control level becomes ISO 2 or 3, and the OEMs then hire me
to make it
an ISO-1 tool for
as little money as possible. At this point, an inexpensive
fix is typically impossible. The higher costs could be avoided
if
tool designers considered
contamination control earlier in the design stage, and used
a contamination control expert to design it.
Occasionally, there is a project that I love. The OEM asks
me to spend a few days with the design engineers at the
start of the
project. Then,
they
call me back three quarters of the way through to make
sure they are on the right track and to discuss any problems that
might arise.
Then
I am
involved
at the end of the project to measure and certify the tool.
This is a good approach to tool design no matter what contamination
levels
are
of concern.
Those projects are great, as everyone is happy with results!
You discuss contamination issues in terms of particles.
What about thin film contamination or airborne molecular
contamination
(AMC)?
(BFK)
Silverman: I think VOC’s (volatile organic compounds) are a huge problem.
I have worked on projects involving VOC’s. However, most of my clients,
particularly the OEM’s that are building tools for wafer fabs and micro-electronics,
do not call me about VOC’s. They continue to be concerned about particles.
There are two reasons for this. The first is that, unlike particles, it is
difficult to see a relationship between VOC’s and yield. The second
reason is that usually the OEMs are not requested to test for VOC’s
in the spec. There is, however, often mention in the specification to use
non out-gassing products (for filter seals, etc.) in the mini-environments.
While process engineers may be concerned about VOC’s, my clients are
usually the product engineers who follow the specifications given to them
by the fab, and these don’t contain VOC specs,
only particle specs.
The pharmaceutical industry is more concerned with
VOC’s,
but have more stringent requirements (e.g. FDA regulations)
than the semiconductor
industry.
In addition to VOC’s, there is concern with other
organic residues, such as polyethylene glycol (PEG),
particularly in
the pharmaceutical
industry. Even if storage containers are eliminated
as a source of residue, PEG contamination
from makeup, hand cream, soap, etc. is a ubiquitous
problem.
(BFK note: Steve is using VOC’s to describe organic
chemicals that can volatilize and contaminate an object.
This is in contrast
to the use
of the term to describe chemicals that promote smog
or chemicals that are found at industrial waste cleanup
sites.)
What do you see as challenges in contamination
control for fabrication facilities of the future,
particularly
for nano
devices? (BFK)
Silverman: The semiconductor industry uses the
goals stated in the ITRS (The International Technology
Roadmap for Semiconductors.)
In
terms of
contamination
control, this roadmap encompasses both AMC and
particles.
The road
map goes out for fifteen years, and has goals
for every year. The roadmap
will take
us into the era of the nanotechnology. The biggest
problem in meeting the contamination control
goals established
in the roadmap is that,
in many
cases, the technology for measurement has not
kept up with the goals.
However, we also need to consider Moore’s Law and the relationship
to the trend toward further miniaturization. According to Moore’s Law,
every 18 months you cut the dimensional metrics by a factor of two, or you
double the memory, speed, etc. Moore’s Law has held true for 40 years
in the semiconductor world. However, currently there is a big debate as to
how long Moore’s Law will hold true.
Currently we are looking at structure that
is a few atoms thick.
Is there a limit? It
is possible that semiconductor
technology based on silicon will have run its
course within ten years, simply due to molecular
constraints
and the laws of physics
where quantum
effects
begin to arise at these small dimensions.
I would answer the question about nanotechnology
with a more basic one: What will my field
look like in five
to
ten years?
We know there will be a paradigm shift.
In the early 1960’s, technology
was based on vacuum tubes; by the 1970’s,
vacuum tubes were relics of the past. We are
accustomed to
thinking in terms
of silicon technology;
but nanotechnology may not use silicon, and
may be organic-based. There are ideas as to
what may replace
silicon, but we are not
certain what
the field
will look like.
Increasingly, as companies adapt to new,
diverse product lines, cleanrooms become “multi-use,” at
least sequentially. Do you have advice
for coping with these changes, short
of a
total remodel? (BFK)
Silverman: I have that exact situation
in a current project. My emphasis is
not multi-use
cleanrooms.
Instead, I advise
my clients
to use
mini-environments and glove boxes.
That way, two assemblers, or product
lines
in close
proximity can be handling totally different
processes with totally different requirements.
That is a real strength of the glove
box or the
mini-environment.
What is the difference between a glove
box and a mini environment? To
me, a glove box
could
be an
enclosed
aqueous parts washer
in a machine
shop.
(BFK)
Silverman: You will never find
a standard glove box in a semiconductor
fabrication
facility,
as all wafers
today
are
handled by robotic
equipment, and glove
handling is not conducive to
the tight contamination control specs
of semiconductors.
Pharma does
not need this expensive
technology
(which
can often add $20,000
or more to the cost of the mini-environment),
as their specifications are not
as tight as those in the semiconductor
industry;
however, in the future,
as pharma specs get tighter,
I think
that pharma will upgrade from
glove boxes
to
a mini-environments
with
automated
tools.
Furthermore, part of the difference
between glove box and mini-environment
is terminology:
Pharma
talks about
glove
boxes; semiconductor
people refer to mini-environments.
They both refer to an enclosure
that separates the
product from people, but there
is no common term even though
they both basically
perform
the same
function.
How often do you see differences
between the “Training” and the “Practice”?
What are the hardest training
practices to get across?
(BFK)
Silverman: In my job, the
devil really is in the details.
The
specification is met
or not
met often
depending upon
the way
the operator or
assembler adheres to the
training in
actual practice.
You can say:
“ Put on a gown exactly this way.”
“
Don’t wear makeup.”
“ Handle the product exactly that way.”
The operators have all been trained; and they all know the appropriate
practice, but I see the rules being
broken all the time. For example, I can walk into any number of cleanrooms
and
see women
wearing makeup.
I see people
changing the documented process.
This is the reason mini-environments have been
widely
adopted: they work!
They separate the product,
the wafer,
from the people.
This separation accounts
for 90% of successful contamination control.
Contact Information:
Steve Silverman
Bartlett Bay Consulting, LLC
(802) 862-4557
steve@bartlettbayconsulting.com
www.bartlettbayconsulting.com
