View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

provided by CERN Document Server

59

PRESSURE MEASUREMENT − TOTAL PRESSURES

Chr. Edelmann
Otto von Guericke University, Magdeburg, Germany

Abstract
Gauges for the measurement of total pressures in the vacuum range using dif-
ferent physical effects are described (with the exception of ionization gauges
that are discussed in a separate chapter). Advantages, disadvantages, tenden-
cies in development, sources of error, and pressure ranges for different appli-
cations are discussed.

1. INTRODUCTION
The term ‘vacuum’ designates a pressure range which extends from pressures somewhat lower than
atmospheric pressure down to pressures lower than 10-14 mbar. Thus it spans a pressure range with more
than 17 orders of magnitude that can be arbitrarily subdivided into ranges of about three to five decades
of pressure. According to DIN 28 400-1 (in agreement with ISO 3529-1) the overall pressure range is
divided into four ranges. Since the beginning of the nineties pressures lower than
10-12 mbar have been defined as extreme high vacuum. In Table 1 this fifth range is in brackets since it is
still not a standardized definition.

Table 1
Pressure ranges

Definition Abbreviation Pressure range

Rough vacuum 1 mbar ≤ p < atm. pressure
Medium vacuum -3
10 mbar ≤ p < 1 mbar
High vacuum hv -7
10 mbar ≤ p <10-3 mbar
Ultra high vacuum uhv (10-12 mbar) ≤ p < 10-7 mbar
(Extreme high vacuum) (xhv) ( p < 10-12 mbar)

Assuming ideal thermodynamic equilibrium we can define the pressure p by Eq. (1):
F
p= (1)
A

(F force acting perpendicularly onto an area A). According to the international system of units the pres-
sure unit is:
1[p] = 1 [F]/ 1 [A] = 1 N/m2 = 1 Pa = 0.01 mbar.

As the ambient atmospheric air and the residual gas in a vacuum chamber both consist of i different gas
components each of them has a partial pressure ppart/i. The partial pressures of all gas components add
linearly to give a total pressure ptotal according to Dalton’s´s law:
ptotal = ∑ p part / i . (2)
i

In future I consider only the total pressure ptotal which I abbreviate by p. The following different physical
effects, depending on the gas pressure, can be used for the measurement of total pressures:
 60

• force effect
• heat transport by heat conductivity of the gas
• heat transport by convection of the gas
• viscosity of the gas
• radiometer effect
• scattering of electrons by collisions with gas particles
• excitation of gas particles by electron impact
• ionization of gas particles by electron impact
• deceleration of ions in an electric field by collisions with gas particles.
Gauges using these different effects for the determination of pressures will now be described.

2. GAUGES USING FORCE EFFECTS FOR PRESSURE MEASUREMENT

2.1 Piston gauges

Because of their complicated structure and operation these gauges are only used for calibration pur-
poses. According to Eq. (1) the force F acting on the area A of a moveable piston placed in a cylinder
can be used for the determination of the pressure. Figure 1 represents the principle. A piston P with a
mass m and a cross section of area A limits a well defined gas volume V. If the pressure in V is increased
with the help of a gas inlet the volume can be restored by an additional weight of mass ∆m. For equilib-
rium we find :
(m + ∆m) g
p= (3)
A

where g is the gravitational acceleration.

p→ 0
chamber

∆m pump
P( m)

V
gas inlet

cylinder

Fig. 1 Principle of a Piston gauge

To avoid the influence of the surrounding atmosphere the cylinder and piston are placed in an ad-
ditional chamber, which is evacuated. To reduce the influence of friction between the cylinder and pis-
ton the piston is made to rotate. With the help of such an apparatus as that represented in Fig. 1 it is
possible to calibrate any gauge indicated in the sketch by G.

2.2 Bourdon gauge

The Bourdon gauge (see Fig. 2a) consists of a bent tube with an elliptic cross section closed at one end
and connected at the other open end to the chamber in which the pressure is to be measured. Pressure
 61

differences between the environment of the gauge and the interior cause forces to act on the two walls of
the tube (Fig. 2b) so that it is bent by an amount that depends on the pressure difference between the
environment and the interior. The bending is transformed by a lever to a pointer whose position can be
calibrated. The importance of this type of gauge is that it is very robust and that it covers a range of
pressure measurement from pressures higher than atmospheric pressure down to rough vacuum (about
10 mbar). The accuracy and reproducibility are relatively poor, so that it is not suitable for precision
measurements, and its usefulness for vacuum measurements is limited.

a) b)
Fig. 2 Bourdon gauge, a) principle, b) distribution of forces

2.3 Diaphragm gauges

If a diaphragm or a bellows separates two regions with different pressures (p1, p2,) the difference ∆p
(∆p = p1 – p2) of these two pressures causes a force that deforms the diaphragm or bellows. There are
many possibilities for measuring this deformation, e.g. mechanically by a lever and a pointer, optically
by a mirror and a light pointer, or electrically by changes of the capacity of a capacitor formed by the
diaphragm and an additional electrode which is usually placed in a region of very low pressure (see
Figs. 3a ... 3c). For precision measurements one side of the diaphragm is evacuated to very low pres-
sure. This is called a reference vacuum. The other side is exposed to the pressure to be measured. The
deformation of the diaphragm depends on, but is not proportional to, the pressure difference [1]. These
days linearization of the pressure vs. deformation reading is mostly performed by electronic circuits.
Thus it is possible to make pressure measurements in a range between some hundred mbar and 10-4
mbar with such a precision that this type of gauge can be used as a secondary standard gauge. The lower
pressure limit is caused by the thermal dilatation that has the same order of magnitude as the deforma-
tion at very low pressures. Some special alloys like stainless steel or special ceramics such as Al2O3 with
high density, are used as materials for the diaphragms. Generally the low pressure in the region of the
reference vacuum is maintained by the use of getters. Frequently the electrodes and the circuits for the
pressure reading are placed in the region of the reference vacuum. Figure 4 shows a diaphragm gauge
with electrical reading. The pressure reading is independent of the gas composition. A pressure range
from atmospheric pressure to 10-4 mbar is covered. The high reliability, simple handling, electrical read-
out that can easily be recorded, and the robust construction are reasons to use it for many practical ap-
plications.

P2

P1 P1 P1
P2
a) b) c)
Fig. 3 Diaphragm gauges with a) a mechanical, b) an electrical and c) an optical display
 62

circuit board spacer block

sensor

centering ring diaphragm

housing
O-ring

Fig. 4 Diaphragm gauge with electrical reading

2.4 Piezoresistive diaphragm gauge

Similar to the diaphragm gauge the piezoresistive gauge consists of a small evacuated capsule closed
with a thin silicon diaphragm at the side exposed to the space whose pressure is to be measured. On one
side of this diaphragm are placed thin film piezoresists produced by evaporation and which are con-
nected to form a bridge circuit. By deformation of the silicon diaphragm the bridge is put out of balance
by an amount depending on the deformation. To avoid destruction of the silicon diaphragm by corrosive
gases there exist models in which a small volume filled with special oil is placed between the silicon
diaphragm and the vacuum space. This volume is closed at the side of the vacuum space whose pressure
is to be measured by a thin stainless-steel diaphragm. The special oil serves as an incompressible pres-
sure transducer.
Depending on the gauge construction the pressure reading covers a range of 0.1 to 200 mbar or
1 to 2000 mbar [2]. The pressure reading is independent of the kind of gas.

3. GAUGES USING THE HYDROSTATIC PRESSURE OF MERCURY OR OTHER NON-

VOLATILE LIQUIDS

3.1 U-tube gauge

These gauges use the hydrostatic pressure of a liquid column for pressure reading. Mostly mercury is
used as the liquid because of its low vapor pressure and its cohesion characteristics.

∆h

Fig. 5 U-tube gauge with one closed glass tube

In the simplest case a gauge of this type consists of a U-shaped glass tube closed at the one end
and connected at the other to the chamber whose pressure is to be measured (Fig. 5). The tube is filled
with mercury so that in the volume between the closed end and the level of the mercury column there is
only the vapor pressure of the mercury and is called Torricelli´s vacuum. The difference in heights of
the mercury levels in the two legs of the U-shaped tube is proportional to the pressure
p = ρg∆h (4)
 63

where ρ is the density of the used liquid, usually mercury, g gravitational acceleration, and ∆h the dif-
ference in heights of the levels of the two columns. Because the lowest measurable difference of heights
of the mercury levels is about 1 mm the lowest pressure detectable with this gauge is in the order of
100 Pa (1 mbar). Because of its uncertainty, its limited range of pressure reading, and the absence of
possibilities for electronic recording, this simple type of gauge has nearly no importance for practical
application today.

3.2 Compression gauge

In contrast to the simple U-shaped gauge, more important is the compression manometer developed by
Herbert McLeod in 1874 [3]. This so called McLeod gauge shown in Figs. 6a and 6b uses mercury as a
liquid piston with which the gas of a well-defined volume V is compressed into a known small volume
VM . By this compression the pressure in the closed volume VM is increased from the unknown low pres-
sure p to the measurable higher pressure p + ρg∆h according to Boyle´s Law :
pV = const. (5)

If the pressure p can be neglected in respect to the hydrostatic pressure ρg∆h, p can be calculated by the
formula:
VM
p = ρg∆h . (6)
V

VM ∆h
V

A
Hg
a) b)
Fig. 6 McLeod gauge

To read the pressure it is necessary to determine the ratio of the volumes V and VM and the differ-
ence ∆h of the heights of the mercury levels. To avoid errors caused by different capillary depression
both the capillary containing the little volume VM (in the capillary) and the comparison capillary parallel
to the connection tube are made from the same piece of glass capillary. The ratio of the two volumes
can be determined exactly before the first use of the gauge and represents a special constant for this
gauge. Thus, the McLeod gauge substitutes the measurement of a pressure by the measurement of only
one length. Therefore it was used for a long time as a primary standard for pressure measurement. To
measure the pressure one has to lower the columns of mercury so that the volume V is connected with
the chamber in which the pressure is to be measured. After reaching the pressure equilibrium one has to
compress isothermally the gas in the volume V up to the volume VM and to determine the pressure in this
volume. For this purpose one has to compare the heights of the mercury columns in the volume VM and
in the comparison capillary. After the measurement one has to lower the mercury level so that the vol-
ume V is connected again with the chamber in which the pressure is to be measured. Thus, the meas-
urement is discontinuous and needs some dexterity. For precision measurements one has to stabilise the
temperature in the surroundings of the gauge. The range of pressure measurement depends on the di-
mensions of the McLeod gauge. The lowest measurable pressures are about (1−0.5) 10-6 mbar. But for
this purpose the volume V has to be at least 1000 cm3 and the volume VM about 1 mm3 or smaller. In
 64

such a gauge one has to handle a lot of mercury and any accident has terrible consequences for the labo-
ratory where the gauge is used. Although the law of Boyle and consequently the pressure reading too do
not depend on the gas composition the McLeod gauge is not able to read correctly pressures of condens-
able vapors because of their condensation during the compression process. Therefore the McLeod gauge
is not convenient for industrial use.

4. MODULATION GAUGE
Differentiating Eq. (4) one finds:
∆V
∆p = − p . (7)
V
This equation explains why periodical changes in V result in periodical changes of p so that ∆p is pro-
portional to p. The proportionality between these two quantities is used in the modulation gauge of Jur-
geit and Hartung [4] shown in Fig. 7. In this gauge a small volume V is periodically changed with the
help of a piezoelectric oscillator (2). The resulting periodical changes in pressure can be detected with
the help of a sensitive capacitor microphone (1) whose signal is proportional to the changes of the os-
cillating pressure. This device is able to read pressures from atmospheric pressure down to 10-6 mbar.
The reading is only slightly influenced by the gas composition as a consequence of the conductance (3)
between the oscillating volume of the gauge and the chamber in which the pressure is to be measured.

Fig. 7 Principle of the the modulation gauge of Jurgeit and Hartung

5. GAUGES USING THE VISCOSITY OF A GAS FOR PRESSURE MEASUREMENT

5.1 The physical basis

According to the kinetic theory of gases the viscosity η is described by Eq. (8) under the assumption
that the equilibrium of the ideal gas is not disturbed:
1
η= ρΛ v (8)
3

with ρ the density of the gas, Λ the mean free path of a gas particle, and v the average value of the ve-
locity of a gas particle. If the mean free path is comparable to the dimensions of the system the viscosity
decreases with lowering of the density (or pressure) of the gas. Thus, the viscosity of a gas can be used
to determine the pressure under the mentioned restriction. One has to bear in mind that the viscosity
depends on the kind of gas.

5.2 Former constructions

To measure the viscosity of a gas one can use the measurement of the damping of any oscillating sys-
tem. This idea is very old. Since the last century different types of gauges which used this principle
were suggested. They contained an oscillating system [6], e.g. a pendulum of any shape (see Fig. 8a−c),
which oscillated after a short excitation and whose damping was measured by noting the reduction of
the amplitude with time. This procedure is not convenient and has the disadvantages that it is discon-
tinuous and is disturbed by every kind of impact or vibration of the housing at which the pendulum is
 65

fixed. Rotating oscillations (see Fig. 8d−e) were also used for the same purpose [7], but they had the
same disadvantages. Therefore, in 1937 Holmes [8] suggested magnetic suspension of the moving sys-
tem that was damped by the viscosity of the gas. The advantage of this equipment is that there is no
mechanical connection between the housing of the gauge and the rotating or oscillating part.

3
3
2 3
2
4 2

5 1

a) b) c)

mirror mirror

B
2d A
d) e)
Fig. 8 Former constructions of gauges:
a) single-string pendulum (2), suspension (3), housing (4), connection to the chamber (5);
b) double-string pendulum (2), suspension (3);
c) double-string pendulum (2) with mirror (1), suspension (3);
d) torsional vibration gauge with one moveable disc between two fixed discs;
e) viscosity gauge with one rotating disc (A) opposite to movable suspended disc (B)

1.3 The spinning-rotor gauge

In 1968 Fremerey [9] developed a gauge with a freely-rotating steel ball for the measurement of the
viscosity of the gas. According to Fig. 9 this ball (1) was suspended by permanent magnets (2) and
magnetic coils (3) whose excitation current was electronically adjusted. The ball was made to rotate by
a special coil. After reaching a desired frequency the excitation was stopped and the remanent magnet-
ism of the ball induced an electric voltage in a coil so that the frequency could be determined electroni-
cally at intervals of time.
4

2
3

1
5

Fig. 9 Spinning rotor gauge (after a data sheet of the Leybold company)
 66

The viscosity and the pressure could be calculated from the decrease of the frequency ν with time from

πρ K r v ν (t )
p= ln (9)
10αt ν (0)

where r is the radius of the ball, ρK its density, ν(t) frequency of the ball at the time t, ν(0) frequency at
the beginning of the measurement, α the accommodation coefficient and, v is the arithmetic average
velocity of the gas particles. The advantage of this gauge is the simple construction of the sensor. It
consists only of a flange (4) connected with a cylindrical tube (5) closed at the end opposite to the
flange and containing the ball. The arrangement of the permanent magnets and the coils is a separate
structure fixed on the tube.
The lower pressure limit is 10-7 mbar or lower, the upper pressure limit is about 10-1 mbar. Due to
the quantities α and v being specific to the gas the pressure reading depends on the composition of the
gas. The precision of this gauge is so excellent that it can be used as a secondary standard for the cali-
bration of other gauges.

1.4 Miniaturization of viscosity gauges

Attempts have been made to miniaturize the dimensions of viscosity gauges by using oscillator quartzes
made for temperature measurement or in wrist watches. Figure 10 shows the construction principle of
the so called quartz friction gauge suggested by Ono and co-workers in 1986 [10, 11]. This equipment
has the form of a tuning fork, but it is much smaller. It is made of quartz by lithographic processes
similar to those used for the fabrication of semiconductor devices. Two specially shaped Au/Cr elec-
trodes are deposited by sputtering on the surface of the tuning fork. The impedance of the capacitor
formed by these two electrodes changes with the frequency and amplitude of the oscillations of this
tuning fork and, in turn, is influenced by the ambient gas pressure. Thus, the change in impedance Z
from the impedance for the resonance condition at very low pressures Z0 depends on the pressure. As an
example, Fig. 11 shows the difference of the impedance (Z – Z0) vs. pressure for a fork length of
2.5 mm.
D
103
in kΩ

quartz crystal
L
102
0
Z–Z

Au/Cr-electrode
101 L=2,5mm

100

D=0,4mm
10−1
D=0,3mm
10−2
10−3 10−2 10−1 100 101 102 103
p in Torr

Fig. 10 Construction principle of the quartz friction gauge Fig. 11 Difference of the impedance (Z – Z0) vs. pressure for a
fork length of 2.5 mm

6. GAUGES USING THE PRESSURE DEPENDENCE OF THE SPECIFIC HEAT

CONDUCTIVITY OF THE GAS

6.1 The physical principle

Neglecting the disturbance of the thermodynamic equilibrium the kinetic theory of gases can be used to
derive Eq. (11a) for the pressure dependence of the specific heat conductivity of a gas λ :
 67

1
λ = Λρ vcV . (11 a)
3

with Λ the mean free path length of a gas particle, ρ the density of the gas, cV the specific heat capacity
of the gas at a constant volume, and v the average value of the velocities of the gas particles. By com-
parison with Eq. (8) one can write for Eq. (11 a) :
λ = ηcV . (11 b)

Considering the disturbance of the thermodynamic equilibrium by temperature gradients Chapman

found a correction factor for Eq. (11 a or 11 b)
λ = 2,52ηcV . (11 c)

Independent of the special shape one can interpret Eq. (11a) in the following manner: The mean free
path length of a gas is reverse proportional to the pressure, and the density is direct proportional to the
pressure. Therefore, the specific heat conductivity of a gas does not depend on the pressure, because
Eqs. (11a−11c) contain the product of ρΛ. This statement is correct for higher pressures in the vacuum
range, if the mean free path length Λ is lower than the geometric dimensions of the vacuum system. But
if − as a consequence of lowering the pressure − the mean free path length reaches values of the same
order of magnitude as the geometric dimensions of the system, the specific heat conductivity is dimin-
ished if the pressure is reduced. This effect can be used for the measurement of the pressure.

1.2 The general construction of a heat-conductivity gauge

Already in 1906 Pirani [12] suggested a gauge which used the pressure dependence of the heat conduc-
tivity for the pressure measurement.
In the simplest case the heat-conductivity gauge consists (Fig. 12) of a thin wire (diameter 2ri)
which is mounted in the axis of a cylindrical tube (diameter 2ro). The cylindrical tube is connected with
the vacuum chamber in which the pressure is to be measured. The thin wire is heated by a constant
electric power Pel. Heat transport to the walls of the tube is caused by:
• conductivity through the electric feedthroughs IW1

• conductivity via the gas IW2

• radiation from gas IW3 .

to the
l
chamber
2r o
2r i

Fig. 12 General construction of a hea- conductivity gauge

At equilibrium:
Pel = IW1 + IW2 + IW3 = ∑I W . (12)
 68

Heat transfer via the electric feedthroughs depends on the temperature of the heated wire, the material,
and the cross section of the feedthroughs. Heat radiation is described by the law of Stefan and Boltz-
mann and depends on the fourth power of the temperatures of the heated wire and its surroundings. Both
the term IW1 and IW3 are independent of gas pressure and nearly constant.
At higher pressures, when the mean free path length Λ is small in comparison to the diameter of
the wire, the specific heat conductivity of the gas is independent of pressure. Thus a constant heat flow
is transported by the gas from the heater to the wall of the gauge. It depends only on the temperatures of
the heater and of the walls and on the geometric structure, but not on the pressure. But if the pressure is
diminished to values at which the mean free path length Λ is in the order of the diameter of the thin
wire, the heat conductivity of the gas decreases. Simultaneously the heat transport from the wire to the
wall is reduced and the temperature of the wire rises if the electric power is constant. At very low pres-
sures the heat flows IW1 + IW3 dominate and the heat flow via the gas can be neglected. Thus, the tem-
perature of the heated wire becomes stable. This behaviour of the heat flow and the temperature of the
heater operated with constant electric power is shown in Fig. 13. The pressure reading is transformed
into the measurement of the heater temperature. This in turn is possible either by the determination of
the electric resistance of the wire with the help of a Wheatstone bridge or by a thermocouple whose
voltage is measured. Another possibility of measurement is to stabilize the temperature of the heater by
variation of the heating power. At low pressure less electric power is necessary to obtain the desired
heater temperature than at higher pressures.

ΣIW
Σ IW, T

10−3 10−2 10−1 100 101 102 103 104

p in Pa

Fig. 13 Heat flow and temperature of a heater with constant electric power

As can be seen in Fig. 13, in the case of operation with constant heating power the pressure is cor-
rectly measurable only between 10-3 mbar and 1 mbar. Above 1 mbar and below 10-3 mbar, the output
vs. pressure curve is so flat that a correct pressure reading is nearly impossible. To extend the range of
pressure measurement to lower pressures one can reduce the term of the heat radiation by lowering the
heater temperature (i.e. lower electric power) and by cooling the tube wall. Thus, it is even possible to
measure with special gauges pressures of about 10-4 mbar.
An extension of the range of higher pressures is also possible. For this purpose the heat transfer
from the heater to the walls by convection can be used. This was shown by Gorski and co-workers in
Poland [14].
The pressure reading depends on the kind of gas and on the accommodation coefficients of the
heated wire and the inner tube wall. Thus, contamination of the tube wall or the heater surface can in-
fluence the pressure reading.

1.3 Miniaturization of heat-conductivity gauges

In principle it is possible to miniaturize the heat conductivity gauges by using technology used in the
production of semiconductor components. For example Huang and Tong [15] have used a silicon chip
with the dimension of only 1.62 x 2.02 mm. The principle of this element is shown in Fig. 14. The
heater consists of a resistor spread over the whole area of the chip. It is fed via a MOS transistor which
regulates the current so that the temperature of the heater is always constant. (The heating increases the
temperature of the whole chip whose temperature is somewhat higher than ambient.) The temperature is
 69

measured by a diode and stabilized by a proper integrated circuit. The voltage of the heater is used for
pressure measurement. Figure 15 shows schematically the characteristics of this gauge. At pressures
lower than 1 mbar the heat conductivity, and above 1 mbar the heat convection, is used to measure the
pressure.
chip with constant temperature 1000

Vout in mV
temperature – 920
heater
sensor Vout + heat
amplifier 840 conductivity
heat
Heat conductivity on the substrate surface convection
760

10−2 100 102

p in Torr
Fig. 14 Principle of heat-conductivity gauge of Huang and Fig. 15 Characteristics of heat-conductivity gauge of Huang
Tong and Tong

Another possibility for miniaturization of the sensor was developed by Badinter and co-workers
[13], who used a very small heater, but external circuits for the measurement. The gauge had a diameter
of 3.6 mm and a height of 3 mm. The heater had a diameter of 6−7 µm and was coated with an isolating
film of 5−7 µm thickness. The range of measurement was between 0.1 and 100 mbar.

7. RADIOMETER-EFFECT GAUGES
The thermal molecular pressure experimentally demonstrated by Crookes in 1873 was used by Knudsen
[16] in 1910 for the measurement of pressures. To explain this effect we consider in Fig. 16 two plane-
parallel plates 1 and 2 with the same area but different temperatures T1 and T2 . The distance d between
the two plates may be small compared with the mean free path length Λ, so that d < Λ/10. Assuming
T1 > T2 (T2 may also be the temperature of the surroundings) gas particles travelling from plate 1 to plate
2 have a higher impact than gas particles coming from the surroundings to the other side of plate 2.
Thus, different pressures seem to act onto the two sides of plate 2. Their difference can be calculated
with the help of the known laws of the kinetic theory of gases by the following equation:

p α 2 − α1 T
∆p = (1 + )( 1 − 1) (13)
2 α1 + α 2 − α1α 2 T2

where α1 and α2 are the accommodation coefficients of the plate 1 and 2 respectively. According to
Fig. 17 one can suspend plate 2 in the manner used for galvanometers. The pressure difference effects
movement of plate 2. The elastic moment of the suspension tries to compensate the momentum caused
by the pressure difference at the two sides of plate 2. Thus, in the equilibrium the torsion of plate 2 indi-
cates the pressure. Figure 18 shows the torsion angle γ vs. pressure p.

1 2 suspension
mirror movable
d frame
T1 T2 A A
b

plate 1 plate 1

cross section A-A

plate 1 plate 1
frame

Fig. 16 Principle of radiometer-effect gauge Fig. 17 Construction of radiometer-effect gauge

 70

10−4 10−3 10−2 10−1 100 101 102

p in Pa

Fig. 18 Torsion angle γ vs. pressure p

According to Eq. 13 this method of pressure measurement is independent of the kind of gas.
Some versions of this equipment have a low-pressure limit of about 10-8 mbar [19]. Because this type of
gauge is very sensitive to vibrations and impact it is not used industrially today.

8. GAUGES USING THE INTERACTION BETWEEN ELECTRONS AND GAS PARTICLES

Collisions between electrons and gas particles have the following effects:
• elastic scattering of the electrons
• inelastic scattering of the electrons connected with dissociation, excitation or ionisation of gas parti-
cles.

8.1 Gauges using the elastic electron scattering for pressure measurement
Gauges of this type can be used where the electrons have energies lower than the ionisation energy. In
this case the probability for excitation is small and the probability for ionisation is zero.
If electrons are accelerated by a low-voltage, elastic collisions with gas particles cause the ap-
pearance of velocity components perpendicular to the direction of the electric field and the reduction of
the accelerating effect in the electric field. In the worst case, the electrons can reach, in a homogeneous
electric field, a constant drift velocity instead of an accelerated movement. Of course, these effects hap-
pen only if the gas pressure is high enough that the mean free path length of the electrons is much
smaller than the distance between the electrodes producing the electric field. These effects can be used
for pressure measurement.
In 1983 Lucas and Goto [18] used the appearance of velocity components of the gas particles per-
pendicular to a homogeneous electric field for pressure measurement. They found that electrons travel-
ling in the direction of a uniform homogeneous electric field diffuse radially to the direction of this
field. Electrons starting from a point source of the cathode arrive at the anode with a mean radial dis-
placement r 2 whose magnitude can be expressed by the formula:

Dr d 2
r2 = 4 (14)
µ U

with Dr the so-called radial-diffusion coefficient, µ the mobility, d the distance between cathode and
anode, and U the voltage between cathode and anode. The quotient Dr / µ is solely a function of U/(pd)
so that the mean radial displacement is a function of the pressure p.
The gauge of Lucas and Goto using this effect for pressure measurement has two anodes facing
the cathode. One has the shape of a circular disc placed at a distance d from the cathode. The other is a
flat ring surrounding concentrically the disc. The distribution of the electron current to these two anodes
having the same potential can be used for pressure measurement. This type of gauge is shown schemati-
cally in Fig. 19.
 71

C A2
A1
A2

I2 I1

UA
Fig. 19 Scheme of Lucas and Goto gauge

The pressure can be measured in a range between 10-3 mbar and about 100 mbar. The curves of the cur-
rent ratio vs. pressure are not linear. Their shape depends on the kind of gas. A linearization of these
curves seems to be possible with the help of a microcomputer.
Another gauge construction was suggested by Edelmann in 1998 [18−21]. The original idea was
to use the drift velocity u of the electrons in the homogeneous electric field for pressure measurement.
Kapzow [22] has shown that the drift velocity u of electrons in a weak electric field depends on the
pressure according to the formula:
eΛ e
u =α E, (15)
mv

where e is the elementary charge of an electron and Λe its mean free path, E the electrical field strength
equal to voltage U/d, m the mass of an electron, and v the (not exactly defined by Kapzov) average ve-
locity of the electrons. The mean free path length of the electrons is inversely proportional to the pres-
sure. Thus it should be possible to determine the gas pressure by the measurement of the drift velocity u.
Originally, it was intended to determine the drift velocity by the measurement of the electron distribu-
tion onto two electrodes. For this purpose the gauge contained, instead of two flat anodes in one plane
perpendicular to the axis between cathode and anode, a flat grid followed by a flat anode sheet as shown
in Figs. 20a–c. Both electrodes are parallel and placed perpendicularly to the axis between the cathode
and anode. They have the same voltage, about 10–20 V positive with respect to the cathode. The ratio of
the electron current to the grid and the electron current to the anode plate can be used to determine the
pressure. With this construction we have measured the pressure in the range between 10-1 and 100 mbar.
The characteristics of this gauge are shown in the Fig. 21. To read the pressure it is possible to use the
ratio of the grid and the anode current, while the cathode current is kept constant, or to keep either the
grid or the anode current constant and to use the current of the other electrode (i.e. anode or grid respec-
tively) for pressure measurement.
cathode C anode A
grid G C G A

IG IA

UA
C G A

a) b) c)
Fig. 20 Drift-velocity gauge
 72

102 UA = 10 V

ratio IG/IA
101

100

10-2 10-1 100 101 102

p [mbar]

Fig. 21 Characteristics of drift-velocity gauge

An attempt to calculate the distribution of the electrons to the two electrodes with the help of the
classical methods used by Barkhausen or Rothe and Kleen for electron tubes does not give correct re-
sults. The reason may be that these methods neglect increasing velocity components perpendicular to
the direction of the homogeneous electric field with increasing pressure.
Thus, for a better calculation of the current distribution only Monte Carlo calculations seem to be
possible. These were performed by Kauert with the aim to optimise and to miniaturize the electrode
structure.
The advantages of these types of gauges are:
• Large range of pressure measurement from 10-3 mbar (or even 10-4 mbar) to about 100 mbar .
• Fast response for pressure changes, much faster than for example diaphragm gauges.
• Possibilities to miniaturize the gauges.
• Simple electronic circuits.
Disadvantages are:
• As in the case of viscosity and thermal-conductivity gauges the pressure reading depends on the kind
of gas.
However the most important problem for this type of gauge is the electron source. Lucas and
Goto used a photo cathode illuminated by ultraviolet light. This is not convenient for practical applica-
tions. We used a directly-heated iridium cathode coated with yttrium oxide. In the rough vacuum range
the lifetime of such a cathode is limited. It is hoped that field emitter arrays or sandwich cathodes will
have a longer lifetime and will be commercially available for this purpose in the near future.

1.2 Gauges using the inelastic interaction between electrons and gas particles
The energy loss of electrons by inelastic collisions with gas particles can cause an excitation of the gas
particles. The excited particles return immediately into the ground state by radiation of electromagnetic
waves. Although the light emission of gas discharges can be used for a rough estimation of the pressure
range, and in some cases even for information about special components of the gas mixture, this effect
is not suitable for exact pressure measurement. But the energy loss of electrons can also effect the ioni-
zation of gas particles. The use of the gas ionization for pressure measurement was discovered in con-
nection with the development of electron tubes at the beginning of this century. Today it is widely used
for pressure measurement in the so called ionization gauges. These gauges are described in another
chapter of these proceedings.
 73

9. SURVEY OF THE OPERATIONAL RANGES OF VACUUM GAUGES

As a short summary in Table 2 are put together the operation ranges of the different gauges for the
measurement of total pressures (ionization gauges are excluded).

Table 2
Short summary of the operational ranges of different gauges for the measurement of total pressures
(x, pressure reading is independent of the kind of gas; o, pressure reading depends on the kind of gas)

Name of the Extreme Ultra-High Vacuum High Vacuum Medium Rough

Range High Vacuum Vacuum
Vac.
log (p/mbar) -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 10
Piston gauges xxx xxx xxx xxx xxx
Bourdon xxx →
gauges
Diaphragm xxx xxx xxx xxx xxx xxx
gauges
Piezoresistive xxx xxx xxx →
diaphragm
gauges
U-Tube Gauge xxx xxx →
McLeod Gauge xxx xxx xxx xxx xxx xxx xxx xxx xxx
Modulation ooo ooo ooo ooo ooo ooo ooo ooo ooo
Gauge
Spinning Rotor ooo ooo ooo ooo ooo ooo
Gauge
Pirani Gauge ooo ooo ooo
Radiometer (xx) xxx xxx xxx xxx xxx xxx xxx
Gauge
Radial Drift ooo ooo
Gauge
Elastic Scat- (?) (?) (?) ooo ooo ooo ooo
tering Gauge

10. ACKNOWLEDGEMENTS
The author is grateful for the careful support in many technical activities of Mrs. U. Lehmann and
Mr. Dipl.-Phys. Hauer. Both were very helpful in accurately preparing the manuscript and the slides for
this lecture.

11. REFERENCES

11.1 General Reviews

S. Dushman, J.M. Lafferty : “SCIENTIFIC FOUNDATION OF VACUUM TECHNIQUE”, John Wiley
and Sons, Inc. New York, London, 1962
J.M. Lafferty (ed.) : “FOUNDATIONS OF VACUUM SCIENCE AND TECHNOLOGY”, John Wiley
& Sons, Inc. New York , Toronto 1997
 74

Chr. Edelmann, H.-G. Schneider (editors) : “Vakuumphysik und –technik”, Akademische Verlagsg e-
sellschaft Geest & Portig K.-G., Leipzig, 1978
Chr. Edelmann: “Vakuumtechnik : Grundlagen und Anwendungen”, Hüthig-Verlag, Heidelberg, 1985
Chr. Edelmann : “Vakuumphysik : Grundlagen, Vakuumerzeugung und –messung, Anwendungen”,
Spektrum Akademischer Verlag Heidelberg, 1998

11.2 Special Papers

[1] B. van Zyl : Check on the Linearity of a Capacitance Diaphragm Manometer between 10-4 and 10-1
Torr, Rev. Sci. Instrum. 47 (1976) 9, 1214
[2] Catalogue of the company LEYBOLD AG, HV 300, Part A1, date of publication: 05/92, pages 18
and 76/77
[3] H. McLeod, Apparatus for Measurement of Low Pressures of Gas, Phil. Mag. 48 (1874), 110
[4] R. Jurgeit, C. Hartung: Vorrichtung zur kontinuierlichen Messung von Gasdrücken.
Wirtschaftspatent der DDR WP G 01 L/197067 Nr. 130850 vom 25 01 1977, erteilt am 3 5 1978
[5] R. Jurgeit, Dissertation, Technische Universität Dresden, 1981
[6] Chr. Edelmann: Stand und Entwicklungstendenzen der Totaldruckmessung in der Vakuumtechnik.
Vakuum-Techn. 34 (1985) 6, 162 - 180
[7] R. Jaeckel: “Kleinste Drucke, ihre Messung und Erzeugung”, Springer-Verlag, Berlin Heidelberg
1952
[8] F. T. Holmes: Rev. Sci. Instr. (1937) , 444
[9] J. K. Fremerey : High Vacuum Friction Manometer, J. Vacuum Sci. Technol. 9 (1972) 108
[10] T. Kobayashi, H. Hojo, M. Ono: Gas Concentration Analysis with a Quartz Friction Vacuum
Gauge. Vacuum 47 (1996) 6/8, 479 – 483
[11] M. Ono, M. Hirata, K. Kokubun, H. Murakami: Quartz Friction Vacuum Gauge for Pressure
Range from 0.001 to 1000 Torr. J. Vac. Sci. Technol. A 4 (1986) 3, 1728 – 1731
[12] M. Pirani. Ber. Dt. Phys. Ges. 4 (1906), 686
[13] E. Ja. Badinter, B.K. Zalewskii, I. G. Starusch: A Miniaturised Thermo-electric Pressure Gauge (in
Russian), Prib. Tekh. Eksp. (1980), 248 – 250
[14] W. Gorski, A. Smiech: Convection Vacuum Gauge for Use Between 10-3 and 760 Torr, Le Vide
No. 169 (1974), 221
[15] J.B. Huang, Q. Y. Tong : Constant Temperature Integrated Vacuum Sensor, Electron Lett. 24
(1988) 23, 1429 – 1430
[16] M. Knudsen: Ein absolutes Manometer. Ann. Phys. 32 (1910) 809 –842
[17] H. Klumb, H. Schwarz : Über ein absolutes Manometer zur Messung niedrigster Drücke, Z. Phys.
122 (1944), 418
[18] J. Lucas, T. Goto: A Pressure Gauge Utilising the Electron Diffusion Process and Microcomputer
Control, Vacuum 34 (1984) 785 – 789
[19] Chr. Edelmann: Vakuum-Meßverfahren und Vakuum-Meßeinrichtung, Deutsches Patentamt,
02.12.1998
[20] Chr. Edelmann: Vakuum-Meßeinrichtung, Deutsches Markenamt, angemeldet 02.12.1998
[21] Chr. Edelmann, R. Kauert, St. Wilfert: Anwendung der Elektronendriftgeschwindigkeit zur
Druckmessung, Frühjahrstagung der Deutschen Phys. Ges. Münster, 1999
[22] N.A. Kapzow: “Elektrische Vorgänge in Gasen und im Vakuum” (German translation from R u-
sian), VEB Deutscher Verlag der Wissenschaften Berlin 1955
75