Mechanical vacuum pumps*

A.D. Chew
BOC Edwards, Crawley, United Kingdom
Abstract
This presentation gives an overview of the technology of contemporary
primary and secondary mechanical vacuum pumps. For reference a brief
history of vacuum and a summary of important and basic vacuum concepts
are first presented.

Introduction

Vacuum? It is just like turning on a tap. We have come to take for granted vacuum and its application,
in all its degrees from vacuum forming, metallurgy, integrated circuitry fabrication to space
simulation. The current provision, the engineering innovations, and global availability bear testament
to the technical and commercial pioneers of the 19th and 20th centuries. The full history itself can be
traced back to the ancient philosophers, however, here we touch briefly on the revolutionary work of
the Natural Philosophers of the period of the Scientific Enlightenment.
There is some debate as to who achieved the first deliberate experimental vacuum (Berti or
Torricelli), however, it is pretty certain that it was in Italy in the 1640s. The concept of a vacuum or
void was a huge discussion amongst philosophical heavyweights across Europe in the 17th century.
Indeed in reasoning against a vacuum (the Aristotolean horror vacuii), the Torricellian void in an
inverted mercury column was said to have been filled with elements from the liquid mercury itself. As
well as creating a vacuum, they also recognised the concept of outgasssinga concept familiar to
many practitioners today.
Galileo and Descartes were party to the debate be it via the concept of the weight of air
(measured as 2.2 g/l compared to a modern value of 1.3 g/l). The work spread from Italy (Baliani) to
Mersenne and Pascal in France. It is worth noting that these intellectual endeavours were patronized
by the great realms of the time both for application to live issues (for example, suction of water in
mines) and also for the reflected glory on the patrons themselves. Science as Natural Philosophy was
truly revered in the 17th century.
Von Guericke raised significant issues with the Torricellian void and combined the glass tube
with his own suction pumps (reported by the Roman engineer Vitruvius) to produced the famous
Magdeburg hemispheres. His problems with the correction of the (degassing) effects of the mercury
column also paved the way for the real gas correction to the ideal gas laws. Along the way his
correction to the mercury column also led to its use for weather prediction (barometry) and in 1660 he
predicted the onset of a heavy thunderstorm.
Pascal had by then joined the debate (as a fully fledged vacuist) as to what exactly could
constitute a void. He led against the Aristotelean view, a victory which was to lead to the development
of the gas laws.

Disclaimer. The author and his employer, BOC Edwards, disclaim any and all liability and any warranty whatsoever relating
to the practice, safety and results of the information, procedures or their applications described herein.

43

A.D. C HEW

1.1

The gas laws

If I have seen further it is by standing on the shoulders of giants a famous quotation from a letter
written by Isaac Newton in 1676. This may have been a way of giving credit to his correspondent,
rival, and contemporary Robert Hooke, but I would like to think that the giants Newton had in mind
would include Robert Boyle who was, amongst many other things, a great vacuum pioneer.
Boyle had a young Hooke as his assistant and circa 1660 they jointly developed an air pump,
arguably the first vacuum-producing apparatus suitable for experiments. These included the
investigation and extinguishing of the ringing of a bell. This is a common vacuum demonstration now
in classrooms but was used by Boyle as evidence for his attack on the plenists (and for theological
applications). This vacuum pump allowed the accurate investigation of the (inverse) relationship
between pressure and volume: P1V1 = P2V2Boyles (or Mariottes) Law.
At this time no temperature scale existed and he could not determine the relationship between
hotness and volume. Later work (1702) by Amontons developed the air thermometerit relied on
the increase in volume of a gas with temperature rather than the increase in volume of a liquid and
gave us Amontons Law: P1T2 = P2T1.
Although Amontons had done the work before him, Charles is credited with the law V1T2 = V2T1
as measured on the Kelvin temperature scale. Gay-Lussac (also famed for his Law of Combined
Volumes) in 1808 made definitive measurements and published results showing that every gas he
tested obeyed this generalization. Daltons work on the law of the partial pressure of gases (1801) and
Avogadros article (1811) stating that at the same temperature and pressure, equal volumes of different
gases contain the same number of molecules would lead to the concept of the mole and Avogadros
number. By 1860 our modern day view of the Ideal Gas Law was in existence and the Maxwell
Boltzmann theory further explained the gas laws in terms kinetics of individual molecules. This
incorporated Bernoullis 1730s kinetic model of tiny gas molecules moving about in otherwise empty
space (a foretelling of the van der Waals equationperhaps the most commonly known of many Real
Gas equations).
Boyles Law was the first of the gas laws and was a truly remarkable display of experimental
procedure with the application of leading-edge contemporary technology. A vacuum pump was the
catalyst and facilitating element as it would prove to be in many other major scientific milestones.
1.2

First industrial applications of vacuum

The development of vacuum technology made many contributions in the 17th to 19th centuries
ranging from evidence to resolve to the philosophical confrontation between plenists and vacuuists, to
the development of the gas laws. By the mid 19th century there seems to have been a major
application of vacuum awaiting discovery. Step forward the humble light bulb.
Davy moved the world on from gas/oil/candle lighting in 1809 by creating an incandescent light
arc-lamp. De la Rue in 1820 followed with an evacuated glass bulb (thence filled with gas) to provide
an incandescent light-bulb though it proved non-economic. From the 1850s work intensified in which
vacuum was used to form the void in which an electric filament was heated to incandescence by an
electric current. Edison is generally attributed with the design in 1879 of the modern-day practical
light-bulb though others (Swan, Weston, Maxim etc.) made many contributions.
Althought they were intended to cause the gas to glow, certain glasses were seen to glow (e.g.,
Geisslers tube) at the end of the tube and/or around the filament. Many investigations of these
cathode rays followed (Lenard and Crookes etc.) and it was a short step for a third electrode to be
used in the vacuum tube to see their effect on these cathode rays. Thomson developed from these
studies his corpuscular theory and discovered the negatively charged electron (1897).

44

M ECHANICAL VACUUM PUMPS

1.3

Vacuum-aided particle discoveries

The understanding of the strange phenomenon of cathode rays presented a great scientific challenge in
the latter stages of the 19th century. In 1871 Crookes suggested that these rays were negatively
charged and in 1879 invented the Crookes vacuum tube (of glass). This was an early form of the much
known 20th century cathode-ray tube and eight years later he performed the famous Maltese cross
experiment. Varley and Crookes in 1871 had suggested that cathode rays were particles and Perrin
proved this in 1895. Positive (canal) rays, or ions, were observed by Goldstein in 1896.
Stoney in 1874 had estimated the charge on the cathode rays (calling them electrons in 1891)
but it was not until 1897 that J.J. Thomson performed his famous series of experiments to demonstrate
the particle or corpuscular, nature of cathode rays. He showed them to be a constituent of an atom
the first subatomic particle to be discovered. It is interesting to note that after Rntgen discovered Xrays in 1895 the first commercial X-ray machine was developed for medical applications in the
following year (by E. Thomson): one of the shortest times to market of any vacuum application!
Rntgen picked up his Nobel prize in physics in 1901 for his discovery of the X-ray.
A fundamental and at least enabling factor in these experiments was the development of the
mean free path theory by Clausius in 1858 without the knowledge that a vacuum is needed to allow
particles to travel unimpeded by collision with gas molecules (as a guide, the mean free path of
nitrogen is 66 cm at 104 mbar). As with today, experimental necessity proved the catalyst for the
development of vacuum generation mechanisms and vice versa new vacuum techniques facilitated
advances in experimental developments. Various pumping techniques for glass vacuum systems had
been developed in parallel to these significant discoveries. These included the Roots pump 1859,
diffusion pumps (Geissler 1855, Topler 1862 and Sprengel 1865), Bunsens 1870 water jet pump and
Dewars 1892 cryogenic (liquid air-cooled charcoal) pump. Fleuss oil piston pump (1892) pump and
Gaedes 1905 mercury sealed rotary vacuum and molecular drag pump of 1912 also played a role.
Subatomic particles were further isolated in 1918 with Rutherfords discovery of the first
nucleon (alpha particle): the proton, and by Chadwick in 1932 of the neutron.
Since then vacuum has continued to be an instrumental tool in driving subatomic studies and
more. Perhaps the most demanding and fundamental studies are still to reveal themselves: the use of
UHV in interferometric experiments for the isolation, capture of gravitational waves. Gravitational
waves were predicted from Einsteins General Theory of Relativity and are disturbances (ripples) in
the curvature of spacetime caused by motions of matter. As these waves pass through matter, their
strength weakens and the wave shrinks and stretches.
Mechanical vacuum pumps developed and used from ~ the mid 20th century to the present day
will be discussed after some basic vacuum concepts.

Basic vacuum concepts

The commonly used and basic vacuum concepts utilize the gas quantities:
number of molecules, N
molar mass, M
gas constant R0
temperature, T
gas mass, w
number of moles, nm
pressure (P) volume (V) units; gas quantity q

45

A.D. C HEW

q = PV ,

(1)

for example, mbar litre/s could be expressed in joules

q = PV = w
PV = nm R0T =

R0T
.
M

(2)

w
R0T .
M

(2a)

q = PV
q = 100 mbar litre
1000 mbar
100 mbar

w
PV = n m R 0T =
R 0T
M

Given PV
Given w

mass can be found

q can be found

Only if we know
M and T

Fig. 1: Gas quantity

Speed, S volume rate

S V .

(3)

Throughput

Q = pressure volume rate (where pressure is constant)

Q = PV = PS .

(4)

RT
Q = PV = w 0 .
M

(5)

Q tells us nothing about pressure and volume rate separatelyonly the product does that.
By speed, manufacturers generally mean the volume flow rate measured under standard
conditions. Generally units are m3/h, l/m or cfm for primary and l/s for secondary pumps but many
other units are used (standards specify between 15C and 25C). Displacement is usually referred by
manufacturers to the swept volume rate D, i.e., the trapped or isolated inlet volume/unit time. The
maximum possible flow rate of the pump, or speed S < D.

46

M ECHANICAL VACUUM PUMPS

2.1

Flow regimes

Knudsen number:

Mean free path

Kn =
d

Characteristic
dimension

Regime

Molecular

Transitional

Continuum

Increasing Pressure
Typically < 103 mbar

Typically > 1 mbar

Fig. 2: Flow regimes

Kn =

(6)

Knudsen number = mean free path/typical dimension.

Regime
Regime
Continuum
Continuum

Molecular
Molecular

<< d
Kn << 1,
moleculemolecule
collisions dominate

Kn >> 1, >> d
moleculesurface
collisions dominate
Fig. 3: Flow regime classification

Definitions:
Kn < 0.01

continuum state

0.01 < Kn < 1

transitional state

Kn > 1

molecular state.

47

A.D. C HEW

Regime
Regime

Molecular
Molecular

Transitional
Transitional

Continuum
Continuum

Type
Type

Laminar
Laminar

Re

Turbulent
Turbulent

Fig. 4: Flow regimes and types

2.2

Conductance, C

Definition

C=

Q
Pu Pd

(7)

Conductance = 1/Resistance
Series

C1

C2

1
1
1
=
+
+"
C n C1 C 2

C1
Parallel

Cn = C1 + C1 + .
C2
Fig 5: Conductance definition

Pumping speed can be combined with a conductance in the same way as conductances in series,
see Fig. 6. N.B. in molecular flow we need to introduce the concept of transmission probability.

48

M ECHANICAL VACUUM PUMPS

Conductance

Pump

Pumping speed can be combined with a

conductance in the same way as
conductances in series

1
1 1
= +
S net S C

Fig. 6: Combination of speed and conductances

Speed, pressure ratio K and conductances are combined as in Fig. 7 so that the zero flow
compression ratio K (i.e. ratio of the upstream to downstream pressures) is given as

Q = Pd S = ( Pu Pd )C
C=

K=

1
1 1
= +
Snet S C

(8)

S
S
or K = 1 + .
K 1
C

Pu

Pu
Pd

Pd

Q
PuSnet = PdS

Snet

S
Snet =
K

From the definition of conductance

K is (zero flow)
compression ratio

Q = PdS = (Pu Pd)C

C=

S
S
or K = 1 +
K 1
C

Fig. 7: Speed, pressure ratio and conductance relationship

The formula for the chamber exhaust: pressure at time t is given by

P(t ) = P(t =0) e

49

S
t
V

(9)

A.D. C HEW

Mechanical pumps

A classification for mechanical pumps of speeds > 1 m3/h is shown in Fig. 8.

Primary pumps exhaust to atmosphere )

Secondary pump exhaust to a backing (primary pump)
Max speeds shown exponent of maximum pump speed 10n (m3/h or l/s)

Wet pumps

Oil sealed rotary vane

Piston
Liquid ring

Primary
Primary
Primary

n = 0 to 3 (m3/h)
n = 0 to 2 (m3/h)
n = 1 to 3 (m3/h)

Primary
Primary or Secondary
Primary
Primary
Primary
Primary
Primary
Primary or Secondary
Secondary

n = 1 to 3 (m3/h)
n = 2 to 5 (m3/h)
n = 1 to 3 (m3/h)
n = 1 to 3 (m3/h)
n = 1 to 3 (m3/h)
n = 1 to 2 (m3/h)
n = 0 to 2 (m3/h)
n = 0 to 2 (l/s)
n = 1 to 4 (l/s)

Dry Pumps

Northey-claw
Roots
Scroll
Screw
Regenerative
Piston
Diaphragm
Drag
Turbomolecular

Fig. 8: Mechanical pump classification scheme

All theses pumps rely on the principle of positive displacement of gas (or vapour) except drag
pumps, which exploit molecular drag and turbomolecular pumps (momentum transfer/capture
technique). Table 1 indicates some of the range of choices involved between wet (oil-filled) and dry
(oil-free) vacuum pumps.
Table 1: Wet and dry pump comparisons
Wet pumps

Dry pumps

Capital cost

Low

High

Oil loss

Can be high at > 1 mbar

Very low

System contamination

Backstream at < 0.1 mbar

Very low

Add on costs

Oil return/filtration

Not necessary

Aggressive process

Not suitable

Resistant

Purge

Sometimes

Almost always

A consideration process involving more decision criteria is shown in Fig. 9.

50

M ECHANICAL VACUUM PUMPS

Roughing time ?
Throughput ?
For process?
Ultimate ?
Leakage/outgassing?
Cleanliness?

Up-time
Service interval?
Running Costs?

$ ?

Choose pump size > requirements

Check: acceptable cost
Pump + foreline still meet requirement
Or re-define requirements/foreline

Fig. 9: Pump choice considerations

Oil-sealed rotary vane pumps

Oil-sealed rotary vane pumps (OSRVs shown in Fig. 10) were first developed in the early 1900s.
Today, the two commonly used oil-sealed pumps are rotary vane and rotary piston pumps. Oil-sealed
rotary vane pumps are often used for low inlet pressures and light gas loads. Oil-sealed rotary piston
pumps are often large and are most often found in high-gas-load, high-inlet-pressure industrial
applications.

EXHAUST
OUTLET

INLET

EXHAUST
VALVE

STATOR

VANE
OIL

ROTOR

Fig. 10: Basic OSRV schematic

The OSRV pumping cycle is shown in Fig. 11.

51

A.D. C HEW

1
2
3
4

Inlet exposed
Trapped volume
Compression
Exhaust

Fig. 11: OSRV pumping cycle

4.1

Functions of oil

The oil in OSRV and piston pumps has several functions:

sealing: the oil surface tension seals the duo-seal and fills the gaps between the vanes, rotors,
and stators;
lubrication: of the bearing areas and blade contact surfaces;
cooling: removes heat from rotors and stators;
protects parts from rust and corrosion: coats and seals surfaces to protect from aggressive gas.
4.2

Single- versus dual-stage pumps

A single-stage pump has one rotor and one set of vanes (ultimate pressure circa 102 mbar). They have
a lower cost where low ultimate vacuum is not required and are used for higher inlet pressures or high
gas loads due to lower compression. A dual-stage pump is most simply two single-stage pumps in
series (ultimate pressure circa 103 mbar or lower). The effective higher compression ratio gives better
ultimate vacuum. These are shown in Fig. 12 and a cutaway is shown in Fig. 13.

Fig. 12: Single- versus dual-stage pumps

52

M ECHANICAL VACUUM PUMPS

Fig. 13: OSRV pump cutaway view

4.3

OSRV gas loads

Pumped gas may contain both permanent gases and vapours, which can condense when compressed.
Condensed vapours may include liquid water and solvents which can mix with pump oil to form an
emulsion. Condensed vapours can limit the ultimate vacuum, cause corrosion, and possibly lead to
pump seizure. Gas ballasting allows vapour pumping without condensation. The basic principle is that
the ballast gas opens the exhaust before compression pressure allows vapours to condense.

Dry pumps clearance mechanisms

Dry pumps do not have oil in the swept pumping volume and use a series of stages with small noncontacting clearances to displace gas/vapours and create compression. Ultimates can be 102 mbar and
lower; higher frequencies giving lower ultimates as they provide reduced (ultimate-limitating) gas
back-leakage. A typical design is shown in Fig.14.

Dry pump: ultimate 0.001 mbar

H e a d p la te
Seal

5 th 4 th

3 rd

2nd 1st

G ear
B e a rin g

Canned
M o to r

C o o lin g

M u lti-L o b e
R o o ts

C law s

Clearance mechanism
Fig. 14: Rootsclaw dry pump schematic

53

R o o ts

A.D. C HEW

5.1

Diaphragm pumps

Figure 15 shows the schematic of a diaphragm pump whose best ultimates are of the order 0.1 mbar.
The crankshaft rotates and the connecting rod pulls the diaphragm down, creating a vacuum in
the chamber. This opens the inlet valve and closes the exhaust valve and the chamber fills with gas. As
the crankshaft continues to rotate, the connecting rod forces the diaphragm to the top of the chamber.
This compresses the gas, opens the outlet valve, and closes the inlet valve. The valves on the inlet and
outlet to the chamber are flapper types, which are operated by pressure. Figure 16 shows a typical
speed curve.

Fig. 15: Diaphragm pump operation schematic

Fig. 16: Vacuubrand/BOC Edwards MD1/XDD1 diaphragm pump speed curve

54

M ECHANICAL VACUUM PUMPS

5.2

Scroll pumps

Eccentrically mounted matching scroll forms move to isolate gas and transport it to an atmospheric
pressure exhaust as in Fig. 17.

Fig. 17: Isolation of gas in scroll pump mechanism

A typical scroll pump speed curve is shown in Fig. 18.

BOCE XDS35i

Fig. 18: BOC Edwards XDS35i speed curve

5.3

Dry piston pumps

The operational principle of dry piston pumps is the same as with oil-filled versions and is shown in
Fig. 19. In the dry version contact can be made with sacrificial coatings.

55

A.D. C HEW

Inlet

Compressed Gas

Exhaust

Trapped Gas

Exhausted Gas

Crankshaft

Fig. 19: Dry piston pump

As the crankshaft rotates, it moves the piston vertically through the cylinder, which traps,
compresses, and then exhausts the gas from the pump. A coating around the outside of the piston
creates the seal between the piston and the cylinder wall.
5.4

Roots/booster pumps

These are non-contacting secondary pumps and the sequence of operation is shown Fig. 20. Rotors
trap a volume of air against the stator body and sweep it around, exhausting the gas at a point 180
from the inlet. Tight clearances between the rotors and the stator are critical to trap and move through
the pump body.
Normally the booster displacement is greater than the backing pump speed therefore large
outletinlet pressure differentials P are generated at high inlet pressure. Unless this P differential is
limited, power demand increases rapidly. Limiting methods, shown in Figs. 21 and 22 include
hydrokinetic drive, inverter motor control and pressure relief valves (outlet to inlet).
INLET
Gas volume
being pumped

2
1

STAGES IN THE OPERATION

OF
A ROOTS MECHANISM

OUTLET

Fig. 20: Booster pump operation

56

M ECHANICAL VACUUM PUMPS

Speed

Hydrokinetic drive
Or bypass valve or
Inverter drive

Direct drive drive

No bypass

0.001

0.01

0.1

Backing pump
Speed

10

Pressure
100

1000

Fig. 21: Mechanical booster limiting methods

Pressure differential
Speed

Increasing
motor
power
Increasing
P
capability

0.001

0.01

0.1

10

Pressure
100

1000

Fig. 22: Mechanical booster P versus power

To achieve speed, ultimate or pump-down requirements, more than one booster may be used.
With boosters in parallel, performance improvement is limited to approximately one decade
improvement in ultimate with modest improvement in pumping speed at high pressures and with high
pumping speed improvement at low pressures. With boosters in series, a lower ultimate is achieved
(limited by outgassing and back-leakage). Also higher pumping speeds at higher pressures are
achieved. This configuration limits the pumping speed at lower pressures. Schematically these are
shown in Fig. 23.

57

A.D. C HEW

Speed

Pressure

B
B

Fig. 23: Schematic of boosters in series and parallel

Turbomolecular pumps

The internal configuration of a bladed and compound/drag turbomolecular (secondary) pump is

shown in Fig. 24. They can generally achieve UHV and exhaust to approximately 0.1 mbar. Different
configurations and compound drag stages can increase throughputs and exhaust pressures significantly
(> 20 mbar).

Fig. 24: Compound turbomolecular pump

58

M ECHANICAL VACUUM PUMPS

6.1

Principle of operation

This is shown in Figs. 25 to 30 below.

Molecules leaving
surface
Surface

1. The direction of the arrow indicates the direction of

travel of the molecule.
2. The length of the arrow indicates the probability
that the molecule will depart in that direction Knudsen
Fig. 25: Principle of operation

Consider (a) text book collision and (b) reality

pII

pII

N
A

P
(a)

(b)

High rotational speeds (> 1000 Hz) tip velocity = molecular thermal
velocities
Fig. 26: Molecular collisions with surfaces

59

A.D. C HEW

Blade speed needs to be of same order as molecular

velocity to influence motion
Turbomolecular
pump blade

Direction of rotation
of blade

Horizontal

In the position shown, there is a higher probability

that the molecules will leave the blade in a
downward direction.
Fig. 27: Molecules leaving a surface

Direction of rotation
of blade

Stator

Fig. 28: Molecules leaving rotors/stators

60

M ECHANICAL VACUUM PUMPS

The blades at the top of the pump have an

open blade structure.
This gives a high pumping speed and low
compression ratio.

The blades at the bottom of the pump have a

closed blade structure.
This gives a low pumping speed and high
compression ratio.

Fig. 29: Open blade structure

Fig. 30: Closed blade structure

The Gaede and Holweck drag stages are shown in Fig. 31. These allow operation to a higher
exhaust pressure.
Holweck

Gaede

Interstage
port

Gas flows from turbo

stages into collection
channel and then into the
first Gaede stage

For ease of explanation

mechanism is shown with
spinning helix. Most
pumps use spinning wall
and stationary helix.
z Spinning helix
z Stationary wall
(pump envelope)

z Spinning discs
z Stationary fingers
(supported by envelope)

Fig. 31: Compound pumping technologies

Conventional bearing turbomolecular pumps have the rotor supported by a combination of

greased/oil-bearings and permanent magnets, as shown in Fig. 32.

Grease

Ceramic/permanent magnet

Fig. 32: Conventional turbomolecular-pump bearing technology

61

A.D. C HEW

Magnetic levitation turbomolecular pumps shown in Fig. 33 are generally more expensive but
are non-contacting with the following advantages:
no bearings to wear and
high reliability
permanently low vibration characteristic with time
no hydrocarbon lubricants present
can be mounted in any orientation
designed to work with (harsh) semiconductor process gases, radiation environments, etc.

3 active axis
Y1

5 active axis

PLANE 1

X1

Y1

PLANE 1

X1

Y2
N S

N S

S N

S N

Cross section
of 5 axis STP

PLANE 2

X2
Z
Fig. 33: Magnetic bearing turbomolecular pumps

A BOC Edwards STPA2203 (nominal speed 2200 l/s) is shown in Fig. 34 with respect to its
throughput performance

Fig. 34: BOC Edwards STPA2203 throughput performance

62

M ECHANICAL VACUUM PUMPS

Other pumps designed for mass spectrometry: split flow/multi-inlet pumps Ref. [1] and pure
drag pumps are also extensively used, especially in machinery for scientific instrumentation.

Regenerative/drag pump

This pump is a single-axis mechanical clearance pump capable of achieving < 106 mbar and
exhausting directly to atmospheric pressure. It is fully described in Ref. [2] and a cross-section shown
in Fig. 35.

Holweck Stages
Rotor
Regenerative
stages

Stator

Fig. 35: Cross-section of the regenerative/drag (EPX) pump

The pump employs a combination of Holweck drag stages (which enable the achievement of the
low ultimate and the molecular pumping speedup to 140 l/s) and regenerative stages which operate
from a few mbar to atmospheric exhaust pressure. Figure 36 shows the pumping speed curves for a
BOC Edwards EPX500 pump which show the typical molecular nature of the pumps performance.
600

Ar
N2

400

He

300

H2

Speed m /hr

500

200

100

0
1.00E-06

1.00E-05

1.00E-04

1.00E-03

1.00E-02

1.00E-01

1.00E+00

Pressure mbar

Fig. 36: EPX500 speed curves for different gases

63

1.00E+01

1.00E+02

1.00E+03

A.D. C HEW

References
[1] A. D. Chew, A. Cameron, D. Goodwin, J. Hamilton, T. Hawley-Jones, P. Meares, J. Pumfrey,
J. Ramsden and D. Steele, Considerations for primary vacuum pumping in mass spectrometry
systems, Spectroscopy 20, January 2005.
[2] A. D. Chew, M. Galtry, R. G. Livesey and I. Stones, Towards the single pump solution: Recent
development in high speed machines for dry vacuum pumping, J. Vac. Sci. Technol. A 23
(2005) 1314.

64