Instrumentation Needs Drive Turbopump Development
Rapidly increasing demand for chemical-analysis
instruments--especially various forms of mass spectroscopy--is pushing
vacuum pump makers to produce cleaner, smaller, and smarter turbopumps.
The following is a manuscript for an article published in R&D
magazine. R&D magazine holds the copyright for the finished
article.
C.G. Masi, Contributing Editor
"The driver right now," says Joe Fabrizio, Director,
Marketing for Instrumentation, Leybold Vacuum U.S.A., Export, Pennsylvania,
"is really drug discovery."
The vehicle Fabrizio refers to being driven is turbopump technology.
Turbopumps (more properly called turbomolecular vacuum pumps) are
provide the operating conditions for equipment in a wide variety
of industries, including semiconductor, aerospace, chemical &
petrochemical, optical and, as Fabrizio points out, pharmaceutical.
While a great many of these applications--representing a huge volume
of turbopump sales--do not involve analytical instrumentation, what
Fabrizio is telling us is that the pressure point for turbopump
technological development right now is in the analytical instrument
area, and specifically analytical instruments serving the drug discovery
and pharmaceuticals industry.
That is not surprising because the pharmaceuticals industry is
undergoing a major technological upheaval.
"What they used to do," Fabrizio recalls, "is they'd
hire a bunch of Ph.D.s and throw them in a room, then, [those chemists]
would come out five years later with a drug.
"What they do now is they start with a general direction to
go in, then pay some guy or a lady eight bucks an hour to use really
fast instruments to run a million samples. They then use sophisticated
computer software to analyze the results of running those million
samples. Out of the million samples, perhaps 100,000 look promising,
so they do more in-depth experiments on those 100,000 samples. They
reduce and reduce until finally they come up with 500 really promising
compounds. Then they give those 500 compounds to the 50 Ph.D.s!"
The instruments used to analyze those millions
of samples are increasingly often automated GC- or LC-MS (see
Table 1) systems such as that shown in Fig. A. While just about
every kind of mass spectroscopy, from quadrupole to time-of-flight,
is represented, they all rely on turbomolecular pumps to create
the conditions under which they can work.
|
Fig. A: Automated mass spectrometers generally rely on
turbomolecular pumps to create the vacuum conditions necessary
for them to run large numbers of samples unattended. Varian
Vacuum Technologies, Lexington, Mass. |
It's no wonder that the bulk of the new-sales volume for turbopumps
comes from this industry. Considering that the analytical instruments
into which those pumps go are undergoing a major technological revolution
as well, it is no wonder that these applications are providing turbopump
manufacturers with major design challenges.
Turbopumped Instrumentation
"Vacuum is not a product," Bill Foley, Marketing Director
at for Varian Vacuum Technologies in Lexington, Massachusetts points
out. "Vacuum is a condition. It's a state that the equipment
has to be in to provide sensitivity, accuracy, and repeatability."
Turbomolecular pumps are a popular means of producing that vacuum
condition because they are clean, rugged, easy to use and relatively
maintenance free.
As Fig. D shows, the classic turbomolecular pump consists of several
stages of what look like fan blades. While they can function like
fan blades at atmospheric pressure or low vacuum, in the high-vacuum
regime where they do most of their work, they actually behave more
like cricket bats or ping-pong paddles to deflect individual gas
molecules toward the outlet (or "high pressure") end of
the pump, where they must be removed from the system by a backing,
or forevacuum, pump.
|
Fig. D: The classic turbomolecular pump uses what look
like fan blades to knock gas molecules from the inlet toward
the outlet. Varian Vacuum Technologies, Lexington, Mass. |
As Fig. B shows, useful turbopumps consist of a rotor carrying
multiple sets of blades rotating at high speed (several tens of
thousands of RPM) separated by sets of stationary blades. Each rotating
set followed by the corresponding stationary set is called a stage.
The stationary blades re-direct the swirling motion molecules pick
up from their collisions with the previous set of rotating blades
into a direction that makes collisions with the next set of rotating
blades more effective.
|
Fig. B: Modern turbopumps consist of several stages of
fanlike rotor and stator blades, often followed by a Holweck
drag-pump stage. Leybold Vacuum U.S.A, Export, Penn. |
Many modern turbopumps follow the fan-like stages with a Holweck
drag-pump stage. The Holweck pump uses viscous forces between a
smooth cylindrical rotor and the gas (now at some-what higher pressure
where viscosity is a useful concept) to drag the gas through spiral
channels machined into a stator toward the turbopump's forevacuum
end.
"Typical turbo pumps exhaust in the middle Torr range,"
says Michael Sears, Marketing Manager, Analytical Instruments, Pfeiffer
Vacuum, Nashua, N. H. "But if you combine a turbo pump and
a drag pump, you get the high compression range [of a fan-type at
low pressure], and you're able to exhaust at higher pressure."
The ability to exhaust at high pressure allows you to pump more
gas out of the system with the same forepump.
The turbopump's ruggedness and versatility is due to this simple
plan. Since the pump operates by mechanical action with only one
moving part, it can be operated (and will even pump gas) at atmospheric
pressure for short periods of time with no damage.
Not surprisingly, different types of analytical instruments present
different technical challenges to their turbopumps.
"You have higher gas flows in MALDI-TOF than in LC-MS in general,"
Tench Forbes, Marketing Manager for the Instrumentation Market Segment
for Europe and the Far East for Varian Vacuum Technologies, Turin,
Italy explains. "You also have multiple-chamber instruments
with TOF and LC-MS, whereas GC-MS typically uses one chamber."
"Typically MALDI-TOF instruments are physically larger and
also far more costly," Roseann Varvaro, Varian's Market Manager
for Instrumentation in North America points out. "So they have
different needs from GC-MS instruments, where cost and space are
more important."
"The vacuum system is one of the most important things in
an MS system," Pheiffer's Sears opines. "It probably is
the most expensive single component in the mass spec other than
maybe the detector."
Thus, cost is a major consideration in developing a turbopump for
GC-MS applications, but it is less a concern (although still a consideration)
in LC-MS applications, especially MALDI-TOF.
One desirable characteristic of turbopumps destined for MS applications
is throughput, or the volume (translated to standard temperature
and pressure) of gas that moves through the system per second. "The
more gas they can bring through," Sears points out, "the
higher the resolution."
Needs for achievable base pressure (that is, the minimum pressure
the turbopump can achieve in the instrument under operating conditions)
also vary from application to application. TOF instruments, which
require gas ions to drift through a relatively long tube as part
of the analysis, have fairly rigorous requirements for base pressure.
IT-MS systems, which are much more compact (see
Sidebar), are much more forgiving of high base pressure.
In summary, the specific requirements for any given application
are likely to be fairly unique to that application. Although trends
are evident within various application groups, the vacuum system
specifications are so intimately coupled to the design details of
the system as a whole, that each application generally requires
a more or less custom design.
This sounds like a recipe for a lot of expensive one-off turbopump
designs, but that is not necessarily the case. Remember that these
pumps are OEM units designed and built to go into new models in
an instrument manufacturer's product line. One reason there is so
much interest in serving the pharmaceuticals market right now is
that a pump designed for a successful product line can generate
significant sales volumes over which to amortize the non-recurring
engineering costs.
The trick, of course, is to create a successful product. While
there is no sure-fire recipe for success, one of the best ways to
avoid disaster is to begin working with suppliers as early as possible.
"We work on being involved up front with the system designers
and giving them ideas on possible performance enhancements for their
vacuum systems," Sears says. "We like to be involved,
perhaps, a year ahead of the prototype phase. At that point, we
can discuss the vacuum needs and the constraints on the vacuum system.
"Our job is then to come in with suggestions. We would present
one or several design possibilities for the turbo pump, the roughing
pump, gauging and a variety of other things. Given advance notice,
we can come in with a solution that allows them to take the next
step in their system."
Multi-Port Turbopumps
As Varian's Forbes pointed out, many LC-MS instruments require
different gas pressures at different locations within the machine.
"They're going from a high pressure region," says Leybold's
Fabrizio, "where they introduce a sample to be analyzed (at
approximately one torr pressure) to the next section, which is normally
between 10-2 and 10-4 torr. That is normally
where they direct the ions toward the detector area. The detector
is yet another chamber, which is usually in the 10-5
to 10-6 torr range."
Standard practice has been to provide a separate pump for each
of these sections. Only the first section is anywhere near the range
at which forepumps operate. All are, however, solidly within turbopump
country. So, this archetypical LC-MS application would call for
three turbopumps operating at different base pressures and gas loads,
along with the associated plumbing, control-lers, etc.
That can get very expensive very fast, and it also will take up
an awful lot of room. It also means a lot of moving parts, which
can cause maintenance and downtime headaches for what Fabrizio calls
"instrument foundries," which are large facilities with
many, many such instruments dedicated to mass-processing of samples
looking for drug candidates--prime customers for the analytical
instrument makers who incorporate turbopumps into their products.
Multi-port turbopumps, such as the one shown in Fig. E, are a much
better alternative for these applications. These units provide two
or more gas inlets that open into different pressure regions of
the pump.
|
Fig. E: Multi-port turbopumps reduce size, complexity and
cost for LC-MS applications. Pfeiffer Vacuum, Nashua, N. H.
|
In any turbomolecular pump, gas enters the main inlet at the base
pressure and is compressed as it passes through the stages in a
serial fashion. So, if you need four different pressures in four
different parts of your MS system, you can plumb the lowest pressure
to the main inlet, where it will see the base pressure, then the
next-highest-pressure chamber to an inlet a few stages farther on,
the next higher pressure to a second intermediate level, and the
roughest vacuum (highest pressure) chamber to, say, just before
the Holweck drag-pump inlet, or even to the forevacuum line.
By carefully selecting how far along the turbopump stack you provide
the inlet for each chamber and adjusting the gas load let into each
chamber, it is possible to tune the system to provide just the right
pressure in each part of the MS system.
The advantages of using a multi-port turbopump in such an application
include lowering size, weight, and cost--just the things system
designers want to minimize. Size and cost are lowered by providing
all the system's pumping needs packaged into one housing containing
one motor, one rotor and one set of electronics. The cost of that
one pump, of course, is likely to be higher than the cost of a conventional
turbopump of equivalent throughput and base pressure. It will, however,
almost surely be considerably lower than two or more such equivalent
pumps.
Multiport turbopumps' disadvantage relative to installing a separate
pump for each vacuum chamber is the plumbing challenge. It's easy
to see the most effective plumbing scheme when you have a separate
pump for each chamber--just hook the pump inlet as directly as possible
into the chamber. It's not so simple when you have one unit serving
several chambers. While the designers of the unit in Fig.
E chose to use a relatively long shaft to put the pumping stages
right where they are needed (minimizing the plumbing between each
chamber and its inlet), the folks who designed the unit in Fig.
G chose to use a somewhat longer high-vacuum line to bring gas into
a more compact pump. These are the types of engineering judgements
that need to be made to fit a turbopump exactly to its application.
|
Fig. G: The design of a multi-port turbopump body should
lend itself to the application. This unit would fit particularly
well where a compact, vertical space is available. Varian Vacuum
Technologies, Lexington, Mass. |
Clean-Bearing Technology
I mentioned "cleanliness" as being one of the turbopumps
signal advantages. To a vacuum engineer, "clean" means
not introducing extraneous volatile materials into the vacuum system.
All materials, including plastics, glasses and even stainless steel,
have a certain probability of vaporizing under vacuum conditions.
The material's vapor pressure, which is a function of temperature,
measures this tendency to evolve vapors.
Obviously, plastics and other organics have much higher vapor pressures
than stainless steel at any given temperature. Turbopumps' cleanliness
stems from the fact that all the mechanical moving parts exposed
to vacuum are made of aluminum or steel, which have negligible vapor
pressure. The electrical components for the motor (and the concomitant
organic plastic insulation materials) can be isolated outside the
vacuum walls. The only source of contamination in a turbopump is
in the bearings, making clean-bearing technology critically important.
Materials normally used to lubricate sliding surfaces tend to be
organic fluids or semifluids. They tend to have vapor pressures
very near or even above atmospheric pressure at normal operating
temperatures. This is why motor oil and grease "smell."
If they didn't evaporate, you couldn't smell them.
Extraneous volatile materials can cause difficulties for MS system
in two ways. They can limit the system's base pressure and they
can create spurious signals in the analysis.
A quantity of volatile material (such as a stick of butter) placed
in a vacuum chamber limits the system's base pressure to its own
vapor pressure. As soon as the chamber pressure goes below the material's
vapor pressure, the material begins evaporating, which pushes the
chamber pressure back up. When the pump removes the vapor, the chamber
pressure again attempts to fall below the material's vapor pressure,
and more material evaporates. The process goes on as long as there
is material in the chamber to evaporate.
The pump, however, can't remove all the vaporized material. Some
of the vapor condenses onto the chamber walls, like dew on grass,
making the chamber's entire inside surface a reservoir of volatile
molecules. Any attempt to pump the chamber to below the material's
vapor pressure is therefore doomed to failure.
Even if the material's vapor pressure is lower than the base pressure
needed for the MS to operate properly, some molecules do evaporate.
These molecules can and will become mixed with the analyte molecules
you are studying, potentially introducing spurious signals into
your results.
Clearly, something must be done to keep contaminant molecules from
backwashing into the MS system. Traditionally, turbopump manufacturers
have gone to great lengths to
- use low-vapor-pressure lubricants, and
- isolate the lubricants from the vacuum by using sealed bearings.
Two additional methods are now being employed in turbopumps intended
for analytical-instrument applications: magnetic bearings and dry
lubricants.
Magnetic Bearings
As any physicist or electrical engineer can tell you, it is impossible
to create a stable levitation system using magnetic forces alone.
If, however, you add an additional non-magnetic force, such stability
can be achieved. By essentially squeezing the turbopump shaft between
two magnetic bearings, engineers learned to use mechanical thrust
to stabilize magnetic bearings for turbopumps decades ago. "Leybold
came out with the first one in the mid seventies," Fabrizio
recalls. "Unfortunately, there was no market."
"It was in the semiconductor industry where the market for
magnetic turbopump bearings really appeared. Truthfully, it's still
limited to the semiconductor industry--to the people who can pay
for it. Magnetic bearings are expensive, but the cost is insignificant
compared to the cost of the rest of the tool."
Magnetic bearings, however, have an obvious advantage as MS users
strive for greater and greater measurement performance. Will turbopump
users be willing to pay the added price for this level of cleanliness?
Will turbopump manufacturers be able to bring the price down enough
to close the cost/benefit gap? We shall see in the next few years.
Dry Lubricants
Another alternative is to eschew viscous-fluid lubricants in favor
of dry lubricants. Manufacturers of magnetic data recording products,
for example, have long used dry-lubricant technology to reduce friction
between read/write heads and hard disk surfaces. That same technology
can help improve turbopumps for analytical instrument applications.
"The bearings in Varian ceramic-bearing turbopumps are permanently
lubricated, so there's no external oil supply of any sort,"
Forbes points out. An additional benefit of this technology is that
the pumps can be mounted at any orientation, horizontally or vertically.
With no liquid lubricants, there is nothing to spill.
Improved Controller Electronics
Not surprisingly, the microelectronics revolution that has made
itself felt everywhere else has also had its effect on the electronics
needed to control turbopumps. Circuit miniaturization has made turbopump
controllers tiny, and microprocessors have made them smart.
"The controller is the brains of the turbopump," Pfeiffer's
Sears points out. "You supply DC power and operation signals,
and you have a fully functional system. It used to have to be mounted
in a behemoth rack mount"
"Instead," says Leybold's Fabrizio, "the controllers
for the new pumps, especially for those intended for the analytical-instrument
market, are little boxes that mount right on the pump. They are
about two inches tall, four inches wide, four inches deep (5 cm
x 10 cm x 10 cm)."
Since they are generally intended for OEM applications where they
are part of a computer-controlled automated system, they can also
incorporate sophisticated internal diagnostics. For example, by
monitoring the pump-motor's current draw, they can get advanced
warning of developing failure conditions.
Rising current draw means the motor is working harder. That can
mean that, say, the bearings are going bad, or a vacuum problem
is making the forevacuum rise. Coupled with other system measurements,
such as the actual forevacuum pressure, this information can help
the host computer monitor the whole system's health.
"What the instrument manufacturers want to do," Fabrizio
says, "is to make the instrument a black box--a no-brainer
for anyone. So none of this functionality is exposed to the user.
The user doesn't know anything about what is going on in the machine."
All they have to think about is the Science that comes out.
Table 1: Mass Spectroscopy Alphabet Soup
Acronym |
Definition |
MALDI |
Matrix-assisted laser-desorption ionization |
TOF |
Time of flight |
GC |
Gas chromatography |
LC |
Liquid chromatography |
IT |
Ion trap |
Back
|