where P is absolute pressure (kilopascal or test. In addition, the system displays the
lbf·in.–2 absolute); Pgage is gage pressure ambient pressure and temperature
(kilopascal or lbf·in.–2); Pbarometer is conditions. Flow measurements, vital to
barometric pressure (kilopascal or lbf·in.–2) leakage verification tests, are also integral
obtained in uncorrected form from local functions. This system accommodates the
weather bureau or read from a precision superimposed leak tests technique or the
barometer and converted to pressure pumpback technique (the mass change
units. Where pressures are measured in verification leak test). The leakage test
other units such as torr, inch of mercury data are represented by a visual display, a
or foot of water, the pressures must be printed record of the raw test data and a
converted consistently either to English concurrent minicomputer calculation of
units or preferably to SI units. After the leakage rates, in several forms.
conversions have been made, the rate of
absolute pressure change can again be Components of Integrated
calculated by use of Eqs. 10 or 11. Leakage Rate Measurement
System
Data Acquisition, Analysis
and Recording Systems for Figure 21 shows a schematic diagram of
Leakage Rate Testing the components and system used in an
integrated leak testing system. Typical
Data recording for large scale pressure leakage rate computations are based on
change leakage tests is made simpler by measurements of the changes in the
sophisticated numerical data acquisition absolute pressure, water vapor pressure
systems. There systems automatically and the dry bulb temperature. The
multiplex the conditioned signals from absolute pressure is measured with a fused
the pressure, temperature, dewpoint and quartz Bourdon tube. The low internal
flow measuring sensors (during the viscosity of fused quartz makes it the most
verification phase of leak testing) at preset perfectly elastic material available. This
automatically timed intervals. Data are type of pressure sensor has no measurable
transmitted through an interface for hysteresis. It also has fast response, high
numerical analysis by computer, recorded resolution and high accuracy. The water
on magnetic tape or disk systems, vapor pressure is measured by use of
displayed by printout or graphical chilled mirror dewpoint sensors and is
recordings and evaluated by error analysis presented to the minicomputer as a
and statistical techniques. dewpoint temperature in degrees celsius
or fahrenheit.
In many cases, this numerical test data
analysis system can analyze the data by The dry bulb temperature is measured
progressive analysis (with least mean by resistance temperature detectors and is
squares fit to straight line approximations also presented to the minicomputer as
of leakage as a function of testing time). digital data. Because the changes in the
Computers provide the fastest and most test parameters are small in magnitude, all
accurate technique for analysis of the input sensors must be capable of high
pressure change leak test data. The data sensitivity, accuracy, repeatability and
can be fed into the computer directly resolution. Similar high accuracy, high
from the acquisition system interface, resolution and reliability are required of
from tape or manually from printer or the electronic networks and digital
recorder readouts. This absolute technique computer analyses.
analysis of leakage rate may be performed
by mass point or leakage rate point-to- Microprocessor Data Acquisition
point, point-to-point cumulative and total and Analyses with Leakage Rate
time statistical analysis techniques. Measurement System
Minicomputer Integrated A microprocessor (minicomputer)
Leakage Rate controls the minicomputer data
Measurement System acquisition and raw data recording
system. The microprocessor system
Integrated leakage rate measurement includes both read-only memory (ROM)
systems are available that include all and random access memory (RAM), a
components from input sensors to scanner system and various interfaces
minicomputer analyses of test data. This with sensor and output system
system will measure and record the components. Digital data for pressure, dry
absolute pressure, the dewpoint bulb temperature and dewpoint
temperature and the dry bulb temperature temperature are presented in ASCII
of the air within the system under leakage (American Standard Code for Information
Interchange) to the computer which then
operates on these raw test data and
calculates the leakage rate (see discussion
below).
188 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Example of Analysis of [∆ P = (42.0 + 14.7)
Data from Pressure Hold
Test of Small Volume Test ∆T
Object
( )− + 89.5 + 460 ⋅ h –1
Table 6 shows leak test data analysis for 42.0 14.7 95.8 + 460 2
pressure hold tests of a small volume test
object with an allowable temperature
corrected pressure loss of 0.5 lbf·in.–2 in = 56.7 − 56.7 549.5 loss in 2 h
2 h. In these tests, corrections for 555.8
variations in barometric pressure were not
required. = 0.6 lbf ⋅ in.–2 loss in 2 h
Analysis of data for the first pressure = 0.3 lbf ⋅ in.–2 ⋅ h–1
hold test on day one by means of Eq. 8
shows that the test object has failed the This leakage rate exceeds the maximum
requirements of the pressure hold test: acceptable temperature corrected pressure
loss of 0.5 lbf·in.–2 in 2 h, so the test
results are not satisfactory. A leak was
located in the pressure zone and the welds
repaired. The pressure hold test was then
repeated four days later with the results
shown in the last two columns of the
table in the left hand column. In this
case, analysis of the pressure hold test
data showed:
FIGURE 21. Information flow diagram for minicomputer controlled integrated leakage rate measurement system using a
microcomputer, dual disk memory and instrument display console.
Containment Console Digital
Quartz manometers
Containment Preset up/down
pressure counter
Analog Digital
Verification air flow Turbine flow meters Microprocessor data
conversion
Resistance
temperature RS-232-C
detectors Input/output
port
Dew point
hygrometers Resistance Analog Digital Analog
temperature
Linear variable detectors (signal
differential and other conditioning)
transducers Scanner
Hygrometer Analog
control conditioning
Digital data encoder
circuitry
Digital
Manual
scanner
Analog Panel meter
Structural Analog
integrity test
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 189
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
∆P = 56.7 − 56.7 550.0 + 56.7 − 56.6 rates of leakage, the compressor must
550.7 operate for a large proportion of the test
time. With low leakage rates, the same
∆T 2 h 2h compressor need operate only
occasionally for relatively short time
= 0.1 lbf ⋅in.–2 loss in 2 h periods. If it is desired to measure the
absolute value of leakage by this
= 0.05 lbf ⋅in.–2 ⋅h–1 technique, the capacity of the compressor
at the test operation pressure should be
This leakage rate is well below the known or be determined in a separate
compressor calibration test.
maximum allowable temperature
corrected pressure loss rate of 0.5 lbf·in.–2 When leak testing by the cyclic
in 2 h, so the test object is now pressurization technique, operators need
two stopwatches and a rapid response
acceptable. pressure gage with a clear scale. Prior to
the first measurements, the compressor is
The calculations in SI would reflect the allowed to charge the system under test to
fact that 1 lbf·in.–2 ≅ 6.9 kPa. its normal operating (or delivery)
pressure. The compressor control valve is
Determining If Pressure then shut off to isolate the test vessel. The
Hold Test is Completed or compressor is allowed to operate under no
Should Be Extended load conditions while the pressure in the
test vessel falls off to a suitable lower
A pressure hold leakage rate test may be pressure limit value, well below the initial
concluded at the end of the required test compressor delivery pressure. When this
period if the magnitude of the pressure pressure limit is reached, one stopwatch is
loss is within (lower than) the specified started and the compressor is put under
allowable rate of pressure loss. If the test load by manual opening of the isolation
results are subject to question, the test can control valve. During this operation step,
be continued over a longer time period to the compressor pumps air or gas into the
increase the reliability of the test data. If test vessel to raise its internal pressure.
the extended test confirms that the actual This first stopwatch is stopped when the
pressure loss rate is in excess of the test pressure has risen to a predetermined
specified allowable limits, the system upper pressure limit.
should be reinspected to detect the
locations of the excess leakage. The When this upper limit is reached, the
system leaks should then be repaired and compressor is cut off by closing the
the system retested to the same isolation valve of the test vessel and the
specifications and procedures. second stopwatch is started. When the
internal pressure of the test vessel or
Leak Testing Techniques system again falls to the original lower
Using Cyclic pressure limit, the second stopwatch is
Repressurization with stopped and reset. Then the first watch is
Compressor started as the compressor is again put
under load to repressurize the test system.
Intermittent operation of a compressor A new cycle of pressurization is initiated
can be used for leakage measurements by and the alternating stopwatch readings for
evaluating the load cycle (duty cycle) pressurization time and leakage time are
when a compressor of known capacity taken. Four or five cycles of
maintains a specified pressure in the repressurization and pressure decay are
system under test. (The duty cycle is the carried out in succession to ensure that
ratio of the time the pump operates, on the compressor is running under constant
time, to the total time testing — on time and reproducible conditions, as required
plus off time for the pump.) With large to obtain accuracy in the leak test
measurements.
TABLE 6. Pressure hold test data. _________D__a_y_O__n_e_________ _________D_a_y__F_o_u_r________
Date: 09:50 11:50 15:30 17:30
Time: 31.9 (89.5) 35.4 (95.8) 32.2 (90.0) 32.6 (90.7)
290 (42.0) 290 (42.0) 290 (42.0) 290 (42.0)
Average surface temperature, °C (°F) 285 (41.4)
Actual test pressure, kPa (lbf·in.–2 gage) 289 (41.9)
Final temperature corrected
4.1 (0.6) 0.7 (0.1)
test pressure, kPa (lbf·in.–2 gage)
Loss in test pressure, kPa (lbf·in.–2)
190 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The cyclic repressurization leak test 2. A rubber dual tubing (two separate
technique is based on the assumption that tubular passages) of length sufficient
no gas is supplied to the system under test to reach from the rubber bags to a leak
during the no load period. With a test control panel. One of the tubular
reciprocating pump compressor unit, passageways connects to the rubber
when the load is removed it is not bags and is used for their
uncommon for the compressor to pressurization. The second tubular
continue to deliver some gas during the passageway extends through the
no load period. This can be avoided by adjacent rubber bag and opens into
ensuring that the compressor is provided the pipeline interior space between the
with a delivery system that enables all gas two rubber bag seals. This tube
in the intercooler to be discharged into transmits the contained natural gas
the atmosphere during the no load pressure to the control panel, where
period. any loss of pressure due to leakage of
gas from the test volume can be
Advantages and Limitations of monitored.
Cyclic Repressurization Leak
Testing 3. A control panel with an inclined water
gage connected to the test volume by
The cyclic repressurization leak testing the rubber tube. This gage is used to
technique has the advantage of requiring measure any variation in gas pressure
only very simple equipment. Its accuracy in the gas line section between the
in leakage measurements is less than the rubber bag seals. A spring gage is used
accuracy of more direct leak testing to indicate the air pressure within the
techniques. It is subject to random errors sealing bags.
caused, for instance, by malfunctioning
compressor valves. Therefore, these valves 4. Connections are provided to a
should be checked for satisfactory pressure pump used to inflate the
operation before starting each cyclic rubber sealing bags and to a suction
pressurization leak test. The cyclic pump that deflates the bags.
pressurization test does not indicate the
volume of the system under test, nor does 5. A steel rod is used to propel the bag
it provide means for leak location. frame and tubing along the inside of
the gas main. The rod and bag frame
When several compressors are have sufficient flexibility to be passed
available, the compressor selected for the through a tape on the gas line and yet
leak tests should be one which if possible have stiffness sufficient to avoid
gives charging times at least as long as the buckling when the apparatus is
leakage times. It is not advisable to pushed along inside the gas main.
operate the compressor under part load
conditions, because its delivery capacity is Procedure for Leak Testing of
rarely determined with the same accuracy Natural Gas Mains with Rubber
for lower loads as for full load. Sealing Bags
Compressors with dead space regulation
have a part load capacity that may differ When testing for leakage in natural gas
between (1) a compressor calibration test mains, the sealing bags described
and (2) a system leakage test, if the previously are inserted into a main
quantity and temperature of the pipeline containing gas under moderate
compressor cooling water are different in distribution pressure. The rubber sealing
these two cases. bags are spaced a set distance apart on the
frame. These bags seal off a portion of the
Localizing Leaks in Low main line when they are inflated with a
Pressure Gas Mains pressure tire pump to 20 to 40 kPa (3 to
6 lbf·in.–2). The gas pressure in the test
Gas utilities use a modification of the volume between the two sealing bags is
pressurizing mode of leakage indicated on the inclined water gage on
measurement to localize leaks in gas the test control panel as soon as the bags
mains. The main is tested, section by are inflated to form pressure seals. When
section, with the leak locator inside the the main gas distribution line is sealed off
pipe. The leak locator consists of the completely by both bags, the pressure in
following parts: the test volume between the bags will
remain constant, as in a pressure hold
1. A flexible frame on which are spaced leak test. Loss of pressure would indicate
two rubber gas bags jointed by a gas leakage through the wall of the
rubber tube. These bags are pressurized distribution pipeline, in the length
to seal off a short section of gas between the two sealing bags.
pipeline for the leakage test.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 191
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 3. Pressure Change Tests for Measuring
Leakage in Evacuated Systems
Introduction to Pressure International System of
Measurements in Units (SI Units) for
Evacuated Systems Vacuum Pressures
By popular usage, atmospheric pressure is The SI unit for pressure is the pascal (Pa)
taken as the upper limit of vacuum. Any and is introduced here as the unit of
pressure less than standard atmospheric pressure in vacuums. Many processes
pressure (101 kPa) is some form of require medium levels of vacuum of the
vacuum. On Earth, vacuum pressure can order of 0.1 to 1 Pa. However, for many
be anything between absolute zero applications such as high altitude
pressure and the barometer reading at the simulation chambers, pressures much
particular location and time. Earlier, the lower than 0.1 Pa are required. Units of
vacuum pressure was measured in inch of millipascal (mPa) or micropascal (µPa) are
mercy (in. Hg) or millimeter of mercury used to describe pressures in this range of
(mm Hg) below atmospheric pressure. A hard vacuum, to avoid negative
vacuum of 28 or 29 in. Hg was considered exponents or powers of ten. The
to be a fairly good vacuum. Now, using SI previously used unit of torr (1 torr =
units, this same vacuum level would be 1 mm Hg) must be multiplied by 133 to
expressed as an absolute pressure of 3 to equal the pressure in pascal. The millitorr
6 kPa, which is 3 to 6 percent of normal is equal to pressure of 133 mPa. Because
sea level atmospheric pressure, 101 kPa the pressure of the standard atmosphere
(1 atm). at sea level is 1.01 × 105 Pa or 101 kPa, it
follows that perfect vacuum would have a
Meaning of Absolute (negative) gage pressure of (–) 101 kPa
Pressure and Gage because the gage pressure in vacuum is
Pressure in Vacuum referred to the standard atmospheric
Systems pressure at sea level.
As suggested earlier, the concept of a Conversions of Vacuum Pressures
vacuum is related to the pressure exerted
by the earth’s atmosphere. Atmospheric from Prior Units to Pascal
pressure indicates the weight of a column
of atmospheric air of unit cross sectional The twentieth century has seen many
area measured at a particular altitude change in the units used to describe
above sea level. With increasing altitude, pressure levels in vacuums. Early
the pressure decreases until, at some investigators described their vacuum
indefinitely great height above the earth’s pressure in terms of millimeter of
surface (where only empty space exists), mercury, or torr, where the atmospheric
the pressure approaches absolute zero. An pressure at standard conditions was taken
enclosure is said to be under vacuum if its as 760 torr. Hard vacuum pressures were
internal pressure is less than that of the later described in terms of micrometer of
surrounding atmosphere. Because of mercury (1 µm is one millionth of a meter
atmospheric pressure changes due to of mercury). Vacuum pressures are
meteorological factors and altitude, the variously expressed in pound per square
numerical value assigned to gage pressure inch absolute pressure (lbf·in.–2 absolute),
in vacuum is referred to atmospheric inch of mercury, torr and the SI unit
pressure under standard conditions at sea pascal. For example, 1 µm Hg = 0.001 torr
level (an absolute pressure of 101 kPa). As = 10–6 m Hg = 133 mPa = 0.133 Pa. The
vacuums were improved, it became pressure of the standard atmosphere is
necessary to provide a scale of absolute then equal to 760 torr. The (negative)
pressures (somewhat analogous to the gage pressure for a perfect vacuum would
scale of absolute temperatures). The then be –760 torr in this system of units.
concept of a perfect vacuum corresponds (An absolute pressure of 1 torr is equal to
to the hypothetical state of zero absolute 133 Pa.) The preferred unit is pascal.
pressure.
Conversion factors relate the various
units used to describe pressures in
evacuated systems, including the pascal,
atmosphere (atm or torr and the
192 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
micrometer of mercury). Figure 22 shows might seem that eventually a pressure of
a scale useful for approximate conversions absolute zero would be obtained. This
of vacuum pressures between SI units of would be true if the only molecules to be
pascal and earlier units of millimeter of removed were those in the gas space.
mercury (equal to torr) for several typical However, other gas sources do exist and
ranges of vacuum pressure. These must be considered. The predominant gas
comparisons may also help personnel to sources are leakage and outgassing.
convert their data into SI units. Leakage is the direct transmission of gas
molecules, driven by the higher external
Limitations on Ultimate pressure, through holes or porosities in
Vacuum Pressure Caused the vacuum chamber wall, in welds or in
by Leakage and the various seals used in the system.
Outgassing Outgassing refers to all forms of gas
coming from the materials in the vacuum
During evacuation of a container, system. It includes gases that are adsorbed
molecules are constantly being removed on the surface, dissolved in the material
by the pumping process. Therefore, it and occluded in gas pockets, as well as
those due to evaporation or
FIGURE 22. Histogram for conversion of vacuum absolute decomposition.
pressures between prior unit of torr and SI unit of pascal
(1 std atm = 100 kPa = 760 torr). The continual addition of gas from
these sources represents the major
1000 133 kPa limitation on the ultimate pressure that
800 100 kPa = 1 atm can be obtained in evacuated systems.
80 Mathematically, the ultimate pressure Pu is
760 60 given by the influx of gas Q divided by
600 the pumping speed S, so that Pu = Q /S.
40 Because the vacuum pump is itself a
400 source of outgassing, it can contribute a
limiting component Pp to the ultimate
200 vacuum pressure. Its effect is frequently
20 included in the prior equation for
ultimate pressure. In this case, Pu = Q /S +
100torr Pp, where the term Q now refers to the
80 kPa influx of gases from all sources except the
60 vacuum pump.
10 kPa
40 8 Even though the pump may be
6 operating at a particular limit pressure for
25.4 torr = 1 in. Hg 4 one type of gas, because of a leak or
20 outgassing, it can still pump other gases
2 to extremely low partial pressure. This is
true because, in molecular flow, all types
10 of gases flow independently of each other.
8 Typically, a gas analysis of an ultrahigh
1 vacuum system operating at a total
pressure of 10 nPa (~1 × 10–10 torr) will
show hydrogen and carbon monoxide as
the residual gases still coming from the
walls of the vacuum system in this
ultrahigh vacuum range. This occurs even
when the partial pressure of the original
nitrogen and oxygen are too low to be
measured.
Pumping Requirements for
High Vacuum Systems
The ideal gas laws apply to ideal gases
even at very low vacuum pressures. They
do not apply, however, to condensable
vapors such as water vapor or refrigerant
gases. The implications of the ideal gas
laws become evident when considering
the effect of reduced pressure on the
volume of a fixed quantity of ideal gas
held at constant temperature. A liter of
gas at standard atmospheric pressure
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 193
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
would increase in volume as pressures are total amount of moisture is very small
lowered in the vacuum region (Table 7). (such as water adsorbed over a small
surface area).
Tremendous multiplying factors come
into existence as the pressure drops in an Ensuring Cleanliness of
evacuated system. The pumping speed in Welded Vessels to Be
cubic meters per second does not increase Evacuated for Leak Testing
as pressure is lowered, so much smaller
masses of gas (fewer gas molecules) are In preparation for leak testing by pressure
removed per unit of time, as system change or helium tests with a tracer probe
pressure drops. or hood, the interior of the system under
test is evacuated. A sensitive vacuum
Residual Gas Molecule pressure gage is then used to measure
Densities in High Vacuum pressure change or a helium mass
Systems spectrometer is used to detect helium
tracer gas that reaches the vacuum pump
Because the mass reduction factor is so input. Joints for high vacuum vessels are
great when evacuating a test system, it far more critical than joints in pressure
might be assumed that after pumping to vessels that also operate under 100 kPa
reach low pressure, there is really nothing (1 atm) of differential pressure.
in the container to affect any work that
may be inserted within it. However, one Microporosity in the weld, entrapped
must consider the number of molecules gases or solids and surface layers that
that remain at various pressures. It may be outgas become major problems with high
recalled that there is a physical vacuum equipment or equipment that
relationship stating that 22.4 L (0.89 ft3) will be evacuated for leak testing.
of any gas will contain 6.023 × 1023 Extremely small defects or inclusions in
molecules at 0 °C (32 °F) and 101.325 kPa. welded joints may not be detectable with
The natural constant, 6.023 × 1023, is the usual nondestructive testing
known as Avogadro’s number. If the gas techniques. The leak testing of the
pressure is now reduced to 0.1 Pa or one evacuated system may be compromised
millionth of its previous value, the 22.4 L because of such small leak and gas
(0.89 ft3) volume of gas still within the sources.
container contains 6 × 1017 molecules.
Even at 1 µPa, some 6 × 1012 molecules For valid leak detection and location by
will still remain in the 22.4 L (0.89 ft3) the tracer probe technique or leakage
volume. This provides a residue of measurement by the hood technique,
3 × 1011 molecules or almost one trillion cleanliness of the test object surfaces and
molecules per liter of volume (one billion the leak testing system is essential. Tracer
per cubic centimeter). To obtain low gas can accumulate in surface dust and oil
ultimate vacuum pressures, one must or grease, including that within leak
reduce the various sources of gas within passageways, possibly causing small leaks
the system being pumped down. Leakage to remain undetected when they are
can be eliminated only by first locating exposed to the tracer probe gas only
each leak and then properly repairing it or briefly. Alternatively, evaporation of
by placing an adequate temporary seal condensed vapors and gases from such
over it. Maintaining cleanliness and contamination layers may cause a
avoiding introduction of moisture into sensitive leak detector system to indicate
the test system before the vacuum leakage when the system is actually
pumpdown are vital. However, where leaktight. The larger the system under
moisture has contaminated the interior test, the more important it is to ensure
volume of a test system, vacuum pumping cleanliness (including weld crevices and
can help to remove the moisture, if the surface discontinuities). The inert gas
tungsten arc welding (GTAW) process
TABLE 7. Gas volume variation with pressure. produces clean welded joints with
minimum permeability to atmospheric or
_G__a_s_P_r_e_s_s_u_r_e_ ___V_o_l_u_m__e_o__f _G_a__s__ tracer gases. The absence of welding flux
minimizes post weld cleaning operations
m3 (ft3) and problems of outgassing from slag
inclusions.
Pa (atm) Description
105 (100) 10 –3 (3.5 × 10–2) atmosphere
103 (10–2) 10 –1 (3.5 × 100)
100 (10–5) 10 2 (3.5 × 103) high vacuum
10–3 (10–8) 10 5 (3.5 × 106) very high vacuum
10–6 (10–11) 10 8 (3.5 × 109)
10–9 (10–14) 10 11 (3.5 × 1012)
194 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Effects of Weld Joint volumes (at the roots of butt welds made
Design on Leak Testing of from two sides of the plate, or fillet welds
Evacuated Vessels with unwelded areas between abutting
plates).
For pressure vessels to be evacuated
during leak testing (and vessels designed Factors Influencing Speed
for vacuum operation), the weld joint of Vacuum Pumping of
design and preparation should avoid Large Volume Systems
trapped volumes or unwelded faying
surface areas that will be exposed to the The pumpdown time or time required for
vacuum side of the joint. Both form evacuation of large vessels and systems
crevices that may hold foreign matter that from atmospheric pressure is highly
can outgas during evacuation or may dependent on the condition of the
provide traps for tracer gases. Because vacuum system, the volume to be
cleaning of such crevices is often evacuated and the pumping speed. Any
impossible, joint design and welding significant amount of water contained in
procedures must eliminate such traps. the system will have a powerful effect on
Welding should be performed from the the time required for pumpdown because
side of the joint that will be evacuated water has a vapor pressure of 2.26 kPa
whenever practical. The under bead often (17 torr) at 20 °C (68 °F). When water is
contains unavoidable microporosity too present within the system to be
small to affect most strength and evacuated, the pressure will not drop
toughness properties of the welded below this value until the bulk of the
structure. However, if exposed to the water has been pumped out. (Drying by
vacuum, these voids could act as trapped evacuation is often a useful way to
volumes. Leakage from this source can be remove water trapped or condensed
avoided by welding the cover (or seal) within pressure vessels, piping and
pass from the side of the pressure components.) Consequently, water or
boundary that will be evacuated. other vaporizing liquids should not be
Figure 23a shows examples of preferred introduced into test systems before leak
joint designs for systems that will be tests that require evacuation, if it can
exposed to high vacuum. Figure 23b possibly be avoided. Evacuation rates
shows undesirable joint designs which attained by mechanical pumps drop
provide dirt traps and create trapped rapidly as the pressure is reduced by
FIGURE 23. Weld joint designs for welded vessels: (a) preferred designs have no crevices or volume traps open to evacuated
side of pressure boundary; (b) undesirable joints trap contamination and tracer gases, which may outgass during evacuation or
leak testing with sensitive mass spectrometer or other vacuum leak detectors.
(a)
(b) T T
T T
T
Legend
= Vacuum side
= Continuous weld
T = Locations of probable gas traps
= Intermittent weld
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 195
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pumping. Gas evolution by evaporation factor of 5, so that the pumpdown time is
of liquids at very low pressures increases estimated as:
rapidly and prolongs the pumping period
required to attain desired vacuums. T = 2.3 × 5 V = 11.5 V
SS
Techniques for Estimating
Time Required for Alternative Technique for
Pumpdown to 10 kPa Estimating Pumpdown
(75 torr) Time to 10 kPa (75 torr)
A technique for approximating the An alternative approximation technique
mechanical pumpdown time for very for estimating pumpdown time of
large systems as given by Guthrie in practical industrial systems with prior
Vacuum Technology7 uses the relation: contamination is also presented by
Guthrie.7 This technique applies for many
(64) T = 2.3 V average industrial systems that may have
S various sources of gas, vapor and leaks
that will require larger pump sizes for any
where T is approximate pumpdown time given pumpdown time. For example, gas
(2.3 time constants) to ten percent of may be trapped on interior surfaces by
initial atmospheric pressure (to about mechanisms such as absorption (which
10 kPa or 75 torr); V is volume of test refers to binding of gas in the interior of
system to be evacuated from atmospheric solid or liquid materials) or adsorption
pressure (100 kPa or 750 torr); S is (which refers to condensation of gas or
pumping speed of evacuation pumps, vapor on the surface of a solid). Despite
volume unit per unit of time. efforts to maintain or restore cleanliness
to the system, there will be variations
Consistent units must be used for each from system to system in the rates of
term in the above equation, such as those outgassing of these trapped gases and
in Table 8. vapors, which will change the required
pumpdown time to achieve specific
Equation 64 indicates the pumpdown vacuum pressures.
time required to reduce pressure to one
tenth of an atmosphere or about 10 kPa. This approximation technique makes
To attain lower vacuum system pressures, use of pumping down curves such as
much more pumping time is required. would apply typically to clean systems of
The term on the right side of Eq. 64 must known interior volume. For typical
be multiplied by the factors in Table 9, for industrial systems with contamination,
various indicated final pressures within leaks or outgassing conditions, the time
the system being pumped down. indicated on the pumpdown curve for
clean systems would be multiplied by a
For example, to evacuate the system to service factor that accounts for the effects
a pressure of only 1 Pa (7.5 mtorr), the of nonideal systems. The service factors to
right side of Eq. 64 is multiplied by a be used for average industrial systems are
listed in Table 10, in terms of the pressure
TABLE 8. Consistent units for pumpdown region to which the system will be
calculation. pumped down.
Time Volume Pumping Speed For example, suppose that a specific
system pumpdown curve shows a
Second Cubic meter Cubic meter/second pumpdown time of 200 min to pump
Second Liter Liter/second from 100 Pa (750 mtorr) to a final
Minute Cubic foot Cubic foot/minute pressure of 10 Pa (75 mtorr). The service
TABLE 10. Service factors for pressure
regions in pumpdown calculations.
TABLE 9. Calculation of approximate _________P_r_e_s_s_u_r_e__R_e_g_i_o_n__________ Service
pumpdown time.
__F_i_n_a_l_P_r_e_s_s_u_r_e__ Multiplying Factor for Pa (torr) Factor
Pa (millitorr) Equation 64 105 to 104 (760 to 100) 1.0
104 to 103 (100 to 10) 1.25
10 100 4 103 to 102 1.5
1 10 5 102 to 101 (10 to 0.5) 2.0
0.1 1 6 101 to 100 (0.5 to 0.05) 3.0
100 to 0 (0.05 to 0.005) 4.0
(0.005 to 0.0002)
196 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
factor corresponding to this final pressure absorption of water vapor or other
range is 2.0. The estimated pumpdown condensable vapors within the interior of
time for this pressure range is then the test volume. The entire interior
obtained by multiplying this pumpdown surfaces were then cleaned with a broom
time of 200 min for an ideal system by a to remove loose dust and dirt to eliminate
factor of 2.0, to obtain an estimated these particles as surfaces on which vapors
pumpdown time of 400 min in this might condense and later outgas.
pressure range for the average industrial
system with contamination or leaks. The technique for calculating
pumpdown time for very large systems is
Comparison of Theoretical used to predict the pumpdown time
and Actual Pumpdown periods ∆t between the initial pressure P1
Curves for Welded Steel and the final pressure P2 in accordance
Tank with the approximation Eq. 65 and 66:8
The following is an example of a (65) ∆t = K V ln P1
theoretical and an actual pumpdown S P2
curve for the annular space of a double
wall vacuum insulated liquified natural (66) ∆t = 2.3 K V log10 P1
gas tank. Figure 24 shows typical pump S P2
curves relating pumping speed to pressure
for a combination of mechanical pumps = K′ V
with booster pumps. For the test to be S
reported here, the pump unit’s
performance curve is typical of several In Eq. 66, K’ = 2.3 K[log10 (P1/P2)]; ∆t is
shown in Fig. 24. Before the pumpdown the pumpdown time between the initial
tests, the annular space welds were
deslagged. The metal surfaces were pressure P1 and the final pressure P2.
examined with a near ultraviolet light Reasonable values for the K and for the K’
(used for fluorescent tracer inspection) to
detect any deposits of hydrocarbons factors are given in Table 11.
(which also fluoresce under ultraviolet
radiation). All deposits detected were then K values cannot be added. However, for
removed with solvent cleaner to reduce calculating the pumpdown time ∆t for a
pressure range that spans two or more of
the pressure ranges listed above, the K’
value to be used is equal to the sum of the
K’ values given for the two or more ranges
covering the pressure difference for which
∆t is to be calculated. For example, if the
FIGURE 24. Curves relating pumping speed to pressure in vacuum chamber for various mechanical pumps with booster pump
units.
600
1 200
Instrument 1
500
Instrument 2 1 000
400
Instrument 3
Speed (L·s –1) 800
Speed (ft3·min–1)
300 Instrument 4 Booster cut 600
in pressure
Booster cut in pressure 2.6 kPa (20 torr)
2 kPa (15 torr)
200 400
100 200
0 0
10–2 10–1 100 101 102 103 104 105
(0.075) (0.75) (7.5) (75) (750)
(7.5 × 10–5) (7.5 × 10–4) (7.5 × 10–3)
Pressure, Pa (torr)
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 197
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pump speed is fairly constant from Equations Used in Analysis of
atmospheric pressure (101 kPa) to a Vacuum Pumpdown Leak Tests
pressure of 100 Pa (1 torr), an estimated ∆t
for the pressure range could be The fundamental response curve for a
determined in one calculation using a K’ vacuum system during pumpdown is
value of 7.3 or (4.0 + 3.3). described by Eq. 67:
The values of K and K’ for computing (67) d P = Q − S P
pumpdown times as listed above apply dt V V
only for the case of clean mild steel tanks.
At pressures below 0.1 Pa (0.001 torr), the where P is pressure in system being
pumpdown times are primarily evacuated; t is time elapsed from start of
determined by outgassing conditions and pumping; S is effective pumping speed of
the relationships of Eqs. 65 and 66 are no vacuum pump; V is volume of system
longer valid. being evacuated; Q is total in-leakage rate
plus outgassing load of test system; dP/dt
Vacuum Pumpdown is time rate of change of pressure. The gas
Technique for Leakage load may be due to leakage, evolution of
Measurements gas from the walls of the evacuated
system or both.
The evacuation pumpdown technique of
leak testing involves the determination In the lower vacuum pressure range
and evaluation of a pressure time response where outgassing has significant effect,
curve for a vacuum test chamber within integration of Eq. 67 leads to the
which the test object is placed for testing. pumpdown response characteristic of
Leakage measurement can be performed Eq. 68:
in either of two ways.
−V 1 − P2 S
1. Determining leakage rate at ln Q
equilibrium pressure attained during (68) t 2 − t1 =
pumpdown. The vacuum test chamber S S
is pumped down to equilibrium 1 − P1 Q
pressure. Test object leakage and
outgassing from the test chamber are Equation 68 describes an exponential
measured and then subtracted from decay curve with a time constant equal to
the value of outgassing measured in a S/V, which becomes asymptotic to an
leakfree system. equilibrium pressure defined by Eq. 69:
2. Deriving an allowable pressure time (69) P = Q
curve for the pumpdown of a system S
under test. Systems deviating from this
relationship are considered to be This ultimate pressure is approached in
leakers. approximately five time constants, when
t2 – t1 = 5(S/V).
With either type of test system, it is
possible to set up an automated leak test Procedure for Pressure Rise
station involving a carousel system. The (Vacuum Retention) Test for
carousel moves the test samples into Leakage Rate
position, pumps them down and
measures the resultant pressures. The The pressure rise test (also called a
biggest difficulty with this type of leak vacuum retention test) is a pressure
test is the false reading produced by change leakage measurement technique
outgassing of dirty samples. performed on a system evacuated below
atmospheric pressure. It can be performed
TABLE 11. Values of pumpdown time estimation factors K on systems at any vacuum level but is
and K’. most effective on systems evacuated to an
absolute pressure (vacuum) in the range
___________P_r_e_s_s_u_r_e_R__a_n_g_e___________ K K’ from 10 to 0.001 Pa (100 to 0.01 mtorr).
Pa (torr) This leakage rate test is performed by
isolating the system under test after it has
101 000 to 2600 (760 to 20) 1.1 4.0 been evacuated to the required (or
2600 to 133 (20 to 1.0) 1.1 3.3 specified) absolute pressure (vacuum).
133 to 13.3 (1 to 0.1) 1.5 3.45 Then the pressure and, when exposed to
13.3 to 6.6 (0.1 to 0.05) 4.0 2.77 ambient weather conditions, the surface
6.6 to 1.3 (0.05 to 0.01) 4.0 6.44 temperature of the system are observed
1.3 to 0.13 (0.01 to 0.001) 4.0 9.21 for a specific time to determine the rate of
pressure rise per unit of time for the
system. Figure 25 shows schematically the
test arrangement and the connections
198 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
FIGURE 25. Arrangement of equipment for pressure rise leakage rate testing of an evacuated
system. Also known as a vacuum retention test, this test measures overall leakage rates and
requires use of a vacuum pumping system and a vacuum gage. For systems exposed to
ambient weather conditions, surface temperature detectors are used to approximate internal
air temperatures in the system. Ambient temperature must be measured in shade, not in
direct sunlight.
Surface thermometer
Boundary of test system
Surface Evacuated
thermometer system
Surface thermometer
Open during test Closed Vacuum pump
Vacuum gage during test system
Gage tube
Optional valve
between the test volume, the vacuum included in the effects described by the
pump system and the instrumentation. General ideal gas law of Eq. 70. For this
reason, in an evacuated system, it is not
Effects of Condensable mathematically realistic to make accurate
Vapors on Vacuum temperature corrections to the final
Retention Leakage Test pressure for pressure data taken at
different temperatures.
As noted earlier in this chapter, the
behavior of vapors in an evacuated system Therefore, to establish a fairly accurate
deviates significantly from the General leakage rate by this pressure change
ideal gas law: technique for an evacuated system
exposed to ambient weather conditions, it
(70) PV = n RT is necessary to compare pressure data at
periods when the temperature is the same
A vapor is the gaseous form of any or nearly the same and the temperature
substance that usually exists in the form trends are in the same direction. For a
of a liquid or solid, such as water vapor. A system enclosed in a temperature
pure liquid in equilibrium with its own controlled building, such as a vacuum
vapor will have two phases (liquid and chamber evacuated to lower absolute
vapor) that coexist at a specific partial pressure ranges, temperature
pressure known as the vapor pressure. measurements are usually not necessary. A
pressure rise test of such an enclosed
Because condensation or evaporation system can be used to determine both the
occurs with changes in temperature, vapor leakage rate and the outgassing rate for
molecules enter or leave the gaseous that system.
phase with any change in temperature.
This changes the number of molecules of Advantages of Pressure
a particular vapor and the partial pressure Rise (Vacuum Retention)
which that vapor exerts in a particular gas Leakage Test Technique
volume. These vapor effects, called
outgassing in a vacuum system, are not The pressure rise leakage rate test is
relatively simple in principle and fairly
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 199
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
easy to perform on smaller test systems. Factors Affecting Leakage
The test is capable of attaining increased Sensitivity of Pressure Rise
leakage sensitivity as the system size or Test Technique
volume decreases. That is, the total
leakage rate that can be measured as a The leakage rate sensitivity of the pressure
pressure rise per unit time becomes rise (or vacuum retention) leakage rate
smaller as the system under test gets test is influenced by five major factors:
smaller in volume. This test technique can
serve as a final test or as a preliminary test 1. absolute pressure attained in the
preceding other leak test techniques, evacuated system, when the test is
depending on the size and configuration performed (this, in turn, affects the
of the system to be leak tested. This resolution of the smallest measurable
quantitative leakage rate test can be used pressure change);
to determine the total leakage rate (in the
form of a pressure rise per unit of time) 2. internal volume of the system to be
through the test boundary of any system tested;
capable of being evacuated.
3. time duration of the leakage rate test;
Limitations of Pressure 4. ambient temperature and weather
Rise (Vacuum Retention)
Leakage Test Technique conditions; and
5. internal surface areas and cleanliness
The sensitivity of the pressure rise leakage
rate test diminishes as the size or volume of the test system.
of the system to be tested increases. Larger
rates of leakage must exist if they are to Each of these factors is discussed next, in
be detected in large volume systems by greater detail.
this test technique. In addition, the
location of unacceptable leakage cannot Effect of Absolute Pressure in
be determined by this test alone. If the Evacuated System Being Tested
actual total leakage rate exceeds the
allowable value, another leak test When vacuum retention leakage rate tests
technique must be used to locate any are performed within the absolute
unacceptable leaks or the numerous small pressure range of 10 to 0.001 Pa (100 to
leaks that might contribute to an 0.01 mtorr) on large systems, the lower
unacceptable high overall rate of leakage. the pressure, the greater the test
Thus, performance of a pressure rise test sensitivity becomes. The limitation on the
on the evacuated annular space of a high pressure end of this range results
double wall vessel, with a resultant total from inability to measure very small
leakage rate indication in excess of that pressure changes resulting from leakage
allowable, will not reveal whether the from large volumes.
unacceptable leakage is in the inner
vessel, in the outer vessel or in a For example, it might be necessary to
combination of both. detect changes of a fraction of a pascal at
2.5 kPa (a few micrometers at 20 torr).
Because of the effect of vapors that do The limitation on the low pressure side is
not obey the general gas laws for ideal the increase in the portion of the pressure
gases, it becomes difficult to determine an change attributable to outgassing. At these
accurate true gas pressure rise per unit of very low absolute pressures, the pressure
time for very large volume systems rise due to actual leakage is small in
exposed to wide temperature variations relation to the pressure rise due to
during the leakage test period. Lowering outgassing. This makes it difficult to
the absolute pressure within the determine the true rate of pressure rise
evacuated vessel in an effort to increase caused by real leakage.
the leakage rate test sensitivity may be
unfeasible because of the vacuum Effect of Volume of Tested System
pumping system limitations. Alternatively,
the rate at which gas can be pumped out The test sensitivity and, in turn, the rate
may be limited by the size of the hole of pressure rise both vary inversely with
(penetration) through which it must be the size or volume of the evacuated system
removed. Trying to increase the test being tested. For example, a leakage rate of
sensitivity by increasing the duration of 5 × 10–3 Pa·m3·s–1 (5 × 10–2 std cm3·s–1) in
the test, in an effort to achieve the ability a 570 m3 (2 × 105 ft3) system would cause
to read a smaller pressure rise per unit a rate of pressure rise of only 0.8 Pa
time more reliably, may prove unrealistic (5.8 mtorr) per day. This same rate of
as costs increase and schedule completion leakage in a 0.3 m3 (10 ft3) system would
is made more difficult. cause a rate of pressure rise of 1.5 kPa
(11.6 torr) per day.
Effect of Duration of Leakage Test
The sensitivity of the leakage rate test
increases directly with the elapsed time
during the test. As the time duration of
200 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
the test increases, the test sensitivity detectable total leakage rate can be
increases. The three factors of absolute computed by Eq. 71. If instead the test
pressure P, system volume V and time sensitivity or total leakage rate Q is
duration t of the pressure rise test are specified or known because of system
related by Eq. 71 and 72: performance requirements, the allowable
pressure rise per unit of time for that total
( )(71) Q = P2 − P1 V leakage rate can be determined by using
t the transposed form of Eq. 71 shown
below as Eq. 73:
where Q is leakage rate (Pa·m3·s–1); P1 is
initial absolute pressure (torr); P2 is final (73) P2 − P1 = Q
pressure (pascal); V is volume of tV
evacuated system under test (cubic meter); or in torr·h–1:
t is time duration of test (second).
( )(72) Q = P2 − P1 V (74) P2 − P1 = 96.6 Q
96.6 t t V
where Q is total leakage rate (std cm3·s–1); Units for variables in Eqs. 73 and 74 are
P1 is initial absolute pressure (torr); P2 is given below Eqs. 71 and 72, respectively.
final absolute pressure (torr); V is volume
of evacuated system (cubic foot); t is time If the rate of pressure rise computed by
duration of test (hour). For other systems Eqs. 73 and 74 is measurable with
of units, the conversion factor of 96.6 will available test equipment at the specified
change. test pressure, the required or specified
leakage rate test sensitivity can be
Effects of Weather and Ambient achieved. If it is not measurable, then an
attainable test sensitivity must be
Temperature Conditions established by Eqs. 71 and 72.
In pressure rise (vacuum retention) tests Example Computation to
of evacuated systems, the greater the Determine Pressure Rise Test
exposure of the system to direct sunlight Feasibility
and the greater the variations in ambient
temperature, the more difficult it becomes As an example of the application of
to determine an accurate pressure rise. Eq. 74, suppose that the performance
Temperature variations lead to specification for a system requires that the
uncontrollable effects on the rate of completed system contain no leakage in
outgassing or condensation of vapors excess of 2 × 10–3 Pa·m3·s–1 (2 × 10–2 std
within the system, which also influence cm3·s–1). This 300 m3 (105 ft3) system can
the pressure variations in the system. be evacuated to an absolute pressure of
1 Pa (or about 10 mtorr) with the
Effects of Internal Surface Area and permanent vacuum pump system. Would
a pressure rise test be a realistic test
Cleanliness of Test System technique for quantitatively verifying that
this system meets the specification
With evacuated systems under pressure requirements? Because Q and V are
rise leak testing conditions, the smaller known, Eq. 73 can be solved as follows in
the internal surface area and the cleaner SI units: (P2 – P1)/t = Q/V =
that surface is, the less the outgassing in (2 × 10–3)/300 = 7 × 10–6 Pa·s–1 or 0.6 Pa
the systems. This reduces the effect on per day. For mixed units, Eq. 74 indicates
pressure change from outgassing due to that:
temperature variations.
Estimating Leakage Test P2 − P1 = 96.6 Q
Sensitivity Attainable in t V
Pressure Rise Tests
= 2 × 10−2
To determine the leakage rate sensitivity 96.6
attainable with a pressure rise (vacuum 104
retention) test, it is necessary to know in
advance the volume (estimated or = 1.93 × 10−2 torr ⋅ h−1
calculated) of the system and the absolute
pressure at which the test must be These requirements can be met by the
performed. If the allowable pressure rise pressure rise (vacuum retention) leakage
per unit of time is known or specified and rate test technique. The time required for
it is realistic for the absolute pressure the test depends on the surrounding
(vacuum) level at which the test is to be temperature conditions. If the system is in
performed, the test sensitivity or a building in a controlled temperature
environment, a test duration of only a
few hours should be adequate. If the
system is exposed to the weather, then a
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 201
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
comparable temperature cycle must be The specification for the pressure rise
experienced. If the weather is cloudy and (vacuum retention) leakage rate test
the temperature is stable, a few hours may required that the test be conducted over a
be adequate. Normally, for an exposed period of 72 h. The allowable pressure rise
system, a temperature cycle of 12, 24 or was 3.3 Pa (25 mtorr) in 72 h. For this
36 h is necessary to achieve the necessary time span, this was a reasonable leakage
reliable comparison data. allowance. For the 650 m3 (2.3 × 104 ft3)
annulus volume and allowable pressure
Example of Pressure Rise Leakage rise rate of 3.3 Pa (25 mtorr), the total
Rate Test of Liquid Hydrogen leakage rate allowable was computed in SI
Vessel units as:
The following example illustrates test Q = 3.3 × 650
conditions and test results for the leakage 72 × 3600
rate of the annular inner space between
concentric inner and outer spheres of a = 8.3 × 10−3 Pa ⋅ m3 ⋅ s−1
double wall vacuum insulated liquid
hydrogen vessel. The outer sphere has a = 8.3 × 10−2 std cm3 ⋅ s−1
15.81 m (51 ft, 10.5 in.) inside diameter
and the inner sphere has an inside The results of the pressure rise test
diameter of 13.9 m (45 ft, 7 in.) and a performed on the annular space of this
wall thickness of about 19 mm (0.75 in.). double wall liquid hydrogen sphere are
The volume of this annular space was shown in the pressure rise test data of
calculated to be about 650 m3 Table 12 and are plotted in the graphs of
(2.3 × 104 ft3). Critical areas of the inner pressure and temperature as a function of
sphere were tested by the more sensitive time during testing in Fig. 26. Pressure
helium tracer probe or hood leak testing levels may be compared at any of the
techniques. nearly equivalent temperature points
during the night time periods marked
FIGURE 26. Graphs showing variations in temperature and absolute pressure of liquid hydrogen sphere annular space during
72 h pressure rise leakage rate test. Arrows with asterisks indicate time periods when temperatures and trends in change of
temperature were comparable. Pressure rise test liquid hydrogen sphere with 13.9 m (45 ft, 7 in.) inside diameter inner tank
and 15.8 m (51 ft, 10.5 in.) inside diameter outer tank.
43 (110)
Temperature, °C (°F) 38 (100) Average shell
32 (90) temperature
27 (80)
21 (70) Ambient
16 (60) temperature
Absolute pressure, Pa (mtorr) 6.7 (50) Absolute
5.3 (40) Pressure
4.0 (30)
2.7 (20)
1.3 (10)
0
600 1000 1400 1800 2200 200 600 1000 1400 1800 2200 200 600 1000 1400 1800 2200 200 600
Real time (h)
04 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72
Elapsed time (h)
202 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
with brackets and footnotes on the Example of Pressure Rise Leakage
elapsed time column of Table 12. These Rate Test of Laboratory Vacuum
time points are also marked by arrows and Chamber
asterisks on the graphs of Fig. 26. The
results for this 52 h time span indicate The following example illustrates test
that the pressure rise on this system could conditions and test results for a pressure
have been a maximum of 250 to 400 mPa rise test of a stainless steel solvent cleaned
(2 to 3 mtorr) in 72 h. This was an vacuum chamber. The purpose of the test
acceptable leakage test rate because it was was to determine the leakage rate and the
much less than the allowable rate of 3 kPa outgassing rate of the chamber. The
(25 torr) in 72 h. The total leakage rate is chamber had an inside diameter of
equivalent to 6.6 × 10–4 to 9.9 × 10–4 590 mm (23.25 in.) and a length of 1.6 m
Pa·m3·s–1 (6.6 × 10–3 to 9.9 × 10–3 std (63 in.). Its volume was calculated to be
cm3·s–1). Because this loss of 250 to about 0.487 m3 or 487 L (17.2 ft3). Its
400 mPa (2 to 3 mtorr) is less than the inside surface area was calculated to be
error in reading of the Pirani or about 4.63 m2 (49.8 ft2). The results of the
thermocouple vacuum pressure gage used pressure rise test with the chamber
for the test, the pressure rise was probably vacuum conditioned for approximately
much less. A longer test period could have 189 h are given in Table 13 and shown
proved this but would have served no graphically in Fig. 27. Figure 27b is an
useful purpose. enlargement of the upper linear portion
of the graph of Fig. 27a, from whose slope
the final leakage rate was determined.
TABLE 12. Test data for pressure rise test of liquid hydrogen sphere.
Real Elapsed ______S_h_e_l_l_T_e_m__p_e_r_a_t_u_r_e_,_°_F_a______ Ambient A_n__n_u_l_u_s_P_r_e_s_s Information and
Time Time (h) Temperature Comments
No. 1 No. 2 No. 3 Average °F Pa (millitorr)
0600 0 59 56 57 57.3 60 2.0 (15) Begin hold test
0800 2
1000 4 65 77 62 68.0 66 2.7 (20)
1200 6
1400 8 72 100 69 80.3 72 3.9 (29) Windy, clear and sunny
1600 10
1800 12 78 112 81 90.3 77 5.1 (38)
2000 14
2200 16b 81 100 100 93.7 80 6.6 (49)
2400 18b
0200 20b 83 97 97 93.0 80 6.9 (52)
0400 22
0600 24 79 87 82 82.7 78 6.3 (47)
0800 26
1000 28 70 69 69 69.3 69 3.7 (28)
1200 30
1400 32 65 65 63 64.3 65 2.7 (20)
1600 34
1800 36 63 62 61 62.0 63 2.4 (18)
2000 38
2200 40b 61 59 59 59.7 61 2.1 (16) Clear and calm
2400 42b
0200 44b 58 57 56 57.0 59 1.9 (14)
0400 46
0600 48 56 53 54 54.3 57 1.5 (11)
0800 50
1000 52 63 81 57 67.0 65 2.4 (18) Clear, calm and sunny
1200 54
1400 56 76 108 69 84.3 77 3.5 (26)
1600 58
1800 60 80 120 85 95.0 79 4.5 (34)
2000 62
2200 64 82 113 96 97.0 80 5.7 (43)
2400 66
0200 68b 84 106 98 96.0 80 6.0 (45)
0400 70b
0600 72b 78 86 80 81.3 78 5.3 (40)
67 70 68 68.3 68 3.5 (26)
63 65 63 63.7 64 2.5 (19)
61 62 61 61.3 62 2.3 (17)
59 60 58 59.0 60 2.1 (16) Clear and calm
57 58 56 57.0 58 1.7 (13)
56 55 54 55.0 57 1.3 (10)
65 87 70 74.0 71 4.0 (30) Clear, calm and sunny
82 120 100 100.7 80 8.0 (60)
85 129 103 105.7 85 10.1 (76)
90 124 106 106.7 87 10.7 (80)
93 120 104 105.7 88 10.7 (80)
87 110 97 98.0 85 10.0 (75)
80 93 86 86.3 80 8.1 (61)
73 78 75 75.3 73 6.0 (45)
70 72 66 69.3 68 3.9 (29)
65 68 62 65.0 65 2.9 (22)
62 63 61 62.0 63 2.5 (19)
59 59 59 59.0 60 2.3 (17) End of 72 h hold test
a. (°F – 32)/1.8 = °C.
b. Data comparison points.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 203
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The total rate of pressure rise due to TABLE 13. Pressure rise leakage rate test of type 487 L
both outgassing and leakage during the stainless steel vacuum chamber. See Fig. 27.
entire 41 h, 44 min (≅ 2500 min =
150 000 s) test period was computed as: Real Time Elapsed _____C_h__a_m_b__e_r_P_r_e_s_s_u_r_e_____
(h : min) Time (h) Pa (torr)
( ) ( )1.64 × 10−2 1.72 × 101
16:36 0 2.9 × 10–4 (2.2 × 10–6)
Total Q = 16:38 0.033 7.2 × 10–4 (5.4 × 10–6)
96.6 × 41.73 16:39 0.05 9.5 × 10–4 (7.1 × 10–6)
16:40 0.067 1.2 × 10–3 (8.7 × 10–6)
= 7.0 × 10−6 Pa ⋅ m3 ⋅ s −1 16:41 0.083 1.3 × 10–3 (1.0 × 10–5)
16:44 0.133 1.9 × 10–3 (1.4 × 10–5)
= 7.0 × 10−5 std cm3 ⋅ s −1 16:47 0.183 2.7 × 10–3 (2.0 × 10–5)
16:53 0.283 4.1 × 10–3 (3.1 × 10–5)
Based on the straight line portion of 16:57 0.35 4.9 × 10–3 (3.7 × 10–5)
the last 23.5 h of the test as shown in the 08:50 16.23 (7.2 × 10–3)
graph of Fig. 27b, the leakage rate Q for 10:50 18.23 1.0 (8.0 × 10–3)
the chamber was computed as: 14:00 21.40 (8.7 × 10–3)
16:50 24.23 1.1 (9.7 × 10–3)
FIGURE 27. Pressure rise leakage rate test of a type 487-L 08:30 39.90 (1.58 × 10–2)
stainless steel vacuum chamber: (a) pressure rise as a 10:20 41.73 1.2 (1.64 × 10–2)
function of time; (b) enlargement of upper portion of curve,
showing rate of pressure rise due to leakage, following 1.3
outgassing of steel vacuum chamber. See Table 13.
2.1
2.2
(a) (10–2) ( ) ( )8.4 × 10−3 1.72 × 101
1.3 × 100 Q=
96.6 × 23.5
Pressure, Pa (torr) 1.3 × 10–1 (10–3)
= 6.4 × 10−6 Pa ⋅ m3 ⋅ s−1
= 6.4 × 10−5 std cm3 ⋅ s−1
Subtracting the leakage rate from the total
rate results in the outgassing rate
computed as:
1.3 × 10–2 (10–4)
1.3 × 10–3 (10–5) Q = 6.0 × 10−7 Pa ⋅ m3 ⋅ s−1
= 6.0 × 10−6 std cm3 ⋅ s−1
1.3 × 10–4 (10–6) 10 20 30 40 50 = 4.6 × 10−6 torr - L ⋅ s−1
0 Elapsed time (h)
This results in outgassing computed in
(b) torr-L·s–1·cm–3 as:
1.5 × 10–2 (2) Outgassing = 4.6 × 10−6
Absolute pressure, Pa (torr) 4.63 × 104
= 1.0 × 10−10
per square meter of surface area. This
agrees very closely with published
outgassing data for degreased stainless
steel with 200 h vacuum conditioning.
Leakage rate
1 × 10–2 (1.3)
5 × 10–3 (0.7) After 10:50
0
4 8 12 16 20 24
Elapsed time (h)
204 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 4. Flow Rate Tests for Measuring Leakage
Rates in Systems near Atmospheric Pressure
Principles of Leakage decrease during pumpdown of a leaktight
Testing by Measurement system.
of Flow Rates
In an alternative leak testing
The flow measurement procedure for procedure, the sealed enclosure can be
leakage testing consists of determining the evacuated and allowed to reach pressure
extent of leakage by measuring the rate of equilibrium with its vacuum pumps. The
flow of gas moving into or out of the rate at which gas is being pumped to
system or component under test. Flow maintain this equilibrium is then
rates can be measured with a flow meter measured to determine the rate of leakage
or by means of pumping at known from the test volume into the enclosure.
volumetric pumping rates to maintain a
fixed system pressure or to compare rates Pumping Technique for
of change of pressure. The flow Measuring Leakage Rate
measurement leakage test procedure can from Pressurized Systems
be roughly separated into two broad
classes of technique: (1) observation and In an alternative pumping technique for
measurement of gas flow rates or volume measuring leakage rates, the test volume
of gas displaced and (2) analysis of effects can be pressurized and the compressor is
of pumping gas during pressurization or then operated only sufficiently to keep
evacuation of systems, on pressure or rates the test system pressure constant. The
of change of pressure. When leak testing leakage rate can then be calculated from
by the flow observation technique, the the volumetric pumping speed (m3·s–1)
amount of leakage is measured. The and the length of time the compressor
system under test is pressurized or must operate to regain a predetermined
evacuated and placed within a sealed system pressure.
enclosure. The enclosure volume is
connected through a flow meter to a Sensitivity of Flow
regulated pressure source. The gas transfer Measurement Leak Testing
by leakage between the system under test Techniques
and its enclosure causes a pressure
difference between the enclosure volume The sensitivity of leakage rate testing by
and the regulated pressure source. The gas flow measurements is relatively low,
transfer between the sealed enclosure and compared to the sensitivity of many other
the reference pressure source is measured leak testing techniques described in this
by flow meters, by movement of a liquid volume. In most cases, the leakage
(slug) indicator in a capillary tube in sensitivity depends on that of the
which the leaking gas is accumulated or
by other techniques. In some cases, the FIGURE 28. Arrangement for leakage rate testing of system
reference pressure may be atmospheric enclosed in a sealed test enclosure connected to a capillary
pressure. Figure 28 shows a leakage testing tube flow meter with an opaque visible liquid indicator slug.
system using a fluid slug indicator of the Leakage from pressurized system into enclosure would cause
amount of gas leakage. indicator slug to move to the right by a displacement
proportional to the volume of gas leakage.
Pumping Technique for
Measuring Leakage Rate System Liquid Connection to
from Evacuated Test under indicator reference volume
Systems or pressure source
test slug (or atmosphere)
In the pumping technique of leakage
testing of evacuated systems, the system Enclosure
under test is evacuated by a vacuum
pump. The rate of system pressure Capillary tube
decrease during pumpdown is then
compared with the rate of pressure
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 205
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
instrument used to measure the flow rate different techniques are used to solve
and is relatively independent of the test individual leak testing problems.
system volume. In a flow observation
technique, leakage rates between 10–3 and Sealed Volume Technique
10–5 Pa·m3·s–1 (10–2 and 10–4 std cm3·s–1) of Leak Testing by Flow
can be detected, depending on the flow Measurements
instrument used. If a sealed system is
being evacuated, flow rates of the order of Figure 28 shows the arrangement of leak
0.1 Pa·m3·s–1 (1 std cm3·s–1) may be testing equipment using the most
observed. (Note that 1 Pa·m3·s–1 is common technique of flow measurement
equivalent to 10 std cm3·s–1.) by observation of the movement of fluid
in a glass capillary tube. The system under
The leakage sensitivity attainable with test is enclosed and sealed within the test
the pumping pressure analysis technique enclosure. The system being tested can be
depends on the size (pumping speed ) of either evacuated or pressurized. It can
the pumps. With evacuated test objects or either be sealed or connected to a source
test systems, leakage sensitivity depends of pressure or of vacuum.
critically on the outgassing within the
system being measured. Care must be taken to ensure that the
leakage being measured is not occurring
Advantages and in the connection to the source of
Limitations of Leak Testing pressure or vacuum. The capillary
by Flow Measurements containing the indicating fluid is attached
to the test enclosure. This type of testing
Flow measurement leak testing procedures can be performed with the capillary fluid
are applicable to a large variety of test indicator connected between the test
systems. The procedures are useful only enclosure and a standard testing volume
for measurement of leakage. They are not on the other end of the capillary. In this
appropriate for locating leaks. They are way, the leak test can be compensated for
used to measure total leakage rates in temperature variations, if both test
small sealed parts. They can be used to enclosure and the comparison volume are
measure total leakage rates in large sealed subject to the same temperature
systems and in systems that can be conditions. Alternatively, the capillary can
pressurized or evacuated. The major be connected between the test enclosure
advantages of leak testing by means of and the atmosphere. For accurate leakage
flow measurements are as follows. measurements and rapid response, the
enclosure containing the system under
1. No special tracer gas is necessary. the test should have a net volume as small as
flow measurement leak testing practical.
procedure is applicable to whatever
fluid is present within the system to One advantage in the construction of
be tested. The test system need not be the sealed volume type of leak testing
placed in any special environment for equipment shown in Fig. 28 is that there
leak testing. Instead, systems may be are no critical, leaktight connections
tested in their normal operating within the enclosure. This is because the
modes. system is operating at atmospheric
pressure. Therefore, although it is possible
2. The cost of the equipment for flow that a leak could exist between the
measurement leak testing is low. enclosure and the atmosphere, leakage
does not occur through this leak because
3. The sensitivity of overall leakage no pressure differential is applied across it.
measurement is independent of Any differences in pressure are
system volume. compensated for by the pressure
transmission through the liquid slug
4. The leakage rate can be measured within the interconnecting capillary tube.
without extensive calibration.
However, the accuracy of leakage Measuring Leakage Rates with Glass
measurement is not very high, as Capillary (Pipette) Tubes
compared with that for many other
techniques. Glass capillary tubes containing a slug of
indicating fluid provide a means for direct
5. When calibration is required, it can be quantitative measurement of leakage rates
readily attained with standard flow or if a record is made of the time required
volume measurement equipment. for the small liquid plug to move a given
distance. Because the cross sectional area
There are two major disadvantages of of the capillary bore is known, the
flow measurement leak testing. volume swept out by the liquid plug
during the measured time interval can be
1. The test sensitivity is low.
2. Flow measurement procedures have
not gained wide recognition.
Flow measurement uses various types of
equipment with little similarity and
206 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
computed. A 1.5 mm (0.06 in.) diameter 4. The fluid should have a low surface
glass capillary tube is used to measure tension so that it can be placed easily
leakage rates in the range from 10–3 to within the bore of the capillary tube.
10–1 Pa·m3·s–1 (10–2 to 100 std cm3·s–1). A
0.5 mm (0.02 in.) glass capillary tube can Alternative Flow
be used to measure smaller leakage rates Measurement Instruments
from 10–5 to 10–3 Pa·m3·s–1 (10–4 to Used in Sealed Volume
10–2 std cm3·s–1). These capillary tubes are Leakage Tests
marked with scales given in convenient
units for computing leakage rates. A The basic principles of sealed volume leak
stopwatch is commonly used for timing testing can be used in numerous ways. For
the movements of the liquid plug within large leaks, flow measuring devices such
the capillary tube. Pipettes used for liquid as a wet type gas meter or a rotameter
measurements provide convenient may be used. These instruments produce
calibrated capillary tubes. accurate leakage rate measurements but
are useful only on very large leaks. For
The upper limit on leakage rates measurements over a wide range of
measurable with capillary tubes is reached leakage rates, the instrument shown in
when the liquid plug moves so fast that Fig. 29 can form a U-tube capable of
timing is difficult. The lower limit on withstanding extremely high pressures.
leakage rate measurement is determined Tubes B and C have different diameters so
by the accuracy desired and is influenced that the proper tube can be selected for
by errors introduced by the resistive and measuring various leakage rates.
inertial forces affecting the movement of
the liquid plug within the capillary tube. When all the valves in Fig. 29 are open
Changes in atmospheric pressure and the test components are pressurized,
(barometric readings) may move the the liquid columns all reach the same
liquid slug in capillary systems with one height. By closing the shutoff valve in the
end open to the earth’s atmosphere. As main line between columns A and B,
the speed of movement of the liquid plug leakage is indicated by upward movement
decreases, these errors are increased. This of fluid in columns B or C. The meter in
causes the leakage measurements to Fig. 29 was designed primarily for
become more inaccurate with slow determining leakage rates in hydraulic
movements of the liquid plug. power systems. The principle of
operations is to displace the leaking fluid
Errors due to starting inertia are with the indicating fluid. This can be
decreased with liquids of lower density. done because there is a pressure loss in
Errors can be reduced, for example, by the leaking component. When the meter
using a water plug about 1 mm (0.04 in.) of Fig. 29 is installed in the hydraulic
long and timing the movement of the power line to the component being
water plug only after it reaches a constant tested, leakage can be measured by the
velocity. If a water plug is used, the error displacement of the separation level
due to the resistive forces of surface between the two different liquids in
tension can be minimized by coating the column A as compared with column B
inside (bore) surface of the clean capillary or C.
tubing with an organosilicon compound.
This coating acts to prevent the water In another type of flow meter, leakage
from wetting the glass. flow in the line between meter and
component under test is measured by the
Mercury is almost impossible to use for
the liquid slug in a glass capillary. FIGURE 29. Connection of delta-vee U-tube manometer for
Mercury has a very high surface tension leakage measurements, with tubes (A,B,C) of different
and it is almost impossible to force it into diameters.
a very small diameter capillary tube bore.
However, there should be negligible gas Shutoff valves
transfer through a mercury plug.
Regulated pressure
An ideal fluid for use as the indicator source
plug in a glass capillary should have the
following characteristics. Test system
1. It should be a fluid in which the A BC
leaking gas is not soluble, so that no
gas transfer by diffusion can occur
through it,.
2. The fluid should not wet the walls of
the tube, so that the surface tension
forces on either end of the plug are
balanced.
3. The fluid should be opaque for easy
visibility and measurement of its
position.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 207
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
displacement of a bellows. The deflection FIGURE 30. Mass flow meter with thermal sensor that
of the pressure difference sensing bellows measures flow through capillary tube: (a) photograph;
system varies the setting of a (b) schematic of components of thermal mass flow
potentiometer. An output electrical signal transducer; (c) temperature distribution under static (no-flow)
from the potentiometer indicates leakage and flowing conditions in flow meter transducer system.
directly in volume units. This bellows
system replaces the observation of (a)
movements of a liquid slug in a capillary
tube. Each of the preceding types of flow
meter will work with liquids as well as
with gases, provided that the indicating
liquid slug is immiscible in the fluid
whose leakage is being measured. This
versatility makes the sealed volume leak
testing techniques extremely useful for
leak testing under operational conditions.
Fast Response Thermopile (b) Direct current voltage source
Mass Flow Meter
Heater
The flow meter (Fig. 30a) comprising a
sensor, electronic circuitry and a shunt Thermocouple 1 Thermocouple 2
measures gas flow rate from 0 to
60 Pa·m3·s–1 (0 to 600 std cm3·s–1). The Meter
shunt causes the flow to divide such that
the flow through the sensor is a precise Heat sink Heat sink
percentage of the flow through the shunt.
The circuit board amplifies the sensor (c)
output linearly to a 0 to 5 V direct current
signal proportional to the flow rate. Tube temperature (relative units) Zero flow
A thermal sensor measures flow Small flow TC-2
through a capillary tube. This flow is a TC-1
fixed percentage of the total flow through
the instrument. This sensor develops an L/2 0 L/2
essentially linear output signal
proportional to flow, which is about Length of tube (relative units)
0.8 mV full scale magnitude (Fig. 30b).
This signal is amplified by the meter
circuitry so that the full scale output is
5.00 V direct current. The output is routed
to interface terminals and to decoding
circuitry in the display.
Measurement of flow rates higher than
60 Pa·m3·s–1 (600 std cm3·s–1) full scale is
achieved by dividing the flow with a fixed
ratio shunting arrangement. The
measuring capillary tube is placed parallel
with one or more dimensionally similar
channels, call laminar flow elements. The
sensor only needs to heat the gas passing
through the capillary tube while retaining
all the mass measuring characteristics.
The fast response of this instrument at
very low rates of air flow permits fast,
accurate leak testing by manual or
automatic means. Table 14 lists
multiplication factors for the air scale
meter indications when this flow meter is
used for gases other than air.
The metal capillary tube of Fig. 30b is
heated uniformly by current from the
transformer. The temperature distribution
is symmetrical about the tube midpoint
with zero flow (Fig. 30b). The external
thermocouples TC1 and TC2 develop equal
but opposing electromotive force outputs
with a symmetrical temperature
208 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
distribution. When air or gas flows gas composition requires only a simple
through the tubing, heat is transferred to multiplier of the air calibration to account
the gas and back again, creating an for the differences in heat capacity. The
asymmetrical temperature distribution flow meter can be used for a wide variety
(Fig. 30c). For constant power input to the of gases during leakage rate testing. The
tube, the differential thermocouple output full scale flow through the flow meter is
voltage is a function of the mass flow rate about 1 Pa·m3·s–1 (10 std cm3·s–1).
and heat capacity of the gas. Changes in
Figure 31 shows typical arrangements
TABLE 14. Multiplication factors for different gases of for leak testing of small items. Figure 31a
mass flow meter air scale.a shows a pneumatic bridge arrangement.
The object to be tested and an identical,
Relative leaktight part used as a reference volume
Conversion Densityc Specific are charged with air at pressures up to
Gas Symbol Factorb (g·L–1) Gravityd 135 kPa (20 lbf·in.–2) gage. The effects of
adiabatic heating or cooling of the air
Acetylene C2H2 0.67 1.09 0.90 during the pressurizing cycle should be
Air 1.00e 1.20 1.00 avoided. The flow meter is then
0.77 0.71 0.59 connected between the unknown and
Ammonia NH3 1.43e 1.66 1.38 reference parts to detect any evidence of
Argon A 0.76 3.25 2.70 leakage (which would allow the pressure
0.88 5.98 4.96 to decrease in the part under test).
Arsine AsH3 0.30 2.51 2.08 Because the adiabatic effects are nearly
Bromine Br2 0.34 2.40 1.99 identical in the reference and the test
Butane C4H10 0.73e 1.84 1.53 parts, the thermopile flow meter quickly
Butene 1 C4H8 1.00e 1.17 0.97 detects the leakage rate without requiring
Carbon dioxide CO2 0.85 2.98 2.47 a waiting period for attainment of full
Carbon monoxide CO 0.45 3.78 3.14 equilibrium in temperatures and
0.52 1.75 1.45 pressures. Leakage testing may also be
Chlorine Cl2 0.50 1.15 0.95 done by a direct inline leak testing
Chlorine trifluoride ClF3 0.56 1.26 1.05 procedure, as sketched in Fig. 31b, but
Cyclopropane C3H6 0.69 1.17 0.97 this test procedure requires a longer time
Diborane B2H6 0.60 1.79 1.49 cycle than the differential flow
Ethane C2H6 0.93 1.58 1.31 measurement technique.
Ethene (ethylene) C2H4 1.43e 0.17 0.14
Ethylene oxide C2H4O 1.03e 0.08 0.07 Orifice Flow Detector with
Fluorine F2 1.01 1.48 1.23 Differential Pressure
Helium He 1.00 1.53 1.27 Transducer
0.85 1.43 1.19
Hydrogen H2 0.31 2.48 2.06 Figure 32a shows a leakage test
Hydrogen chloride HCl 1.39 3.49 2.90 instrument system that uses an orifice to
0.69e 0.68 0.56 convert flow across the orifice element
Hydrogen fluoride HF 1.38 0.84 0.70 into a pressure differential sensed by the
1.00 1.24 1.03 differential capacitance sensor (see also
Hydrogen sulfide H2S 1.02e 1.17 0.97 Fig. 14a). The orifice (which produces a
Isobutane C4H10 0.75 1.85 1.54 pressure loss when air flows through it) is
Krypton Kr 0.97e 1.33 1.10 connected in series with the air supply
0.15 2.83 2.35 line to the item under test, as shown in
Methane CH4 0.22 3.18 2.64 Fig. 32). This system is used with
Neon Ne 0.79 1.53 1.27 automatic flow and leakage testers
0.32e 1.89 1.57 providing fully automatic cycling and
Nitric oxide NO 0.36 5.93 4.92 accept/reject test indicators and output
0.36e 5.13 4.26 signals. Leakage sensitivity and
Nitrogen N2 0.42 4.59 3.81 stabilization time are both programmable.
0.48 3.65 3.04
Nitrous oxide N2O 0.43e 3.65 3.03 A compensation network provides a
0.22e 6.99 5.80 programmed electronic time base signal to
Oxygen O2 0.68 1.33 1.10 match the dynamic characteristics of
0.70 2.72 2.26 short time cycle flow measurements.
Pentaborane B5H9 0.28 6.43 5.34 Figure 32a is a photograph showing
0.23 8.22 6.82 instrument connections to the differential
n-Pentane C5H12 0.23 14.65 12.16 pressure transducer (capacitance gage) and
0.80 0.76 0.63 the orifice or flow restriction element (in
Phosphine PH3 1.37 5.54 4.60 this example, a short length of tubing).
Typical ranges vary from 0.05 to 250 L·s–1
Propane C3H8 (0.002 to 9.0 ft3·min–1). The dynamic
range is indicated to be 50:1 for operation
Refrigerant-11 CCl3F with laminar flow pressure loss elements.
Refrigerant-12 CCl2F
Refrigerant-13 CClF3
Refrigerant-14 CF4
Refrigerant-22 CHCIF2
Refrigerant-114 CClF2
Silane SiH4
Sulfur dioxide SO2
Sulfur hexafluoride SF6
Tungsten hexafluoride WF6
Uranium hexafluoride UF6
Water vapor H2O
Xenon Xe
a. No corrections or compensations for temperature or pressure of gas
required.
b. Multiply air scale by these conversion factors.
c. Density in grams per liter at 20 °C (70 °F) and 100 kPa (1 atm).
d. Specific gravity (air = 1.00).
e. Empirical data; other data is theoretical. Example: Flow meter NALL-1K,
0–1000 std cm3·s–1 in air would be 1000 × 1.43 = 1430 std cm3·s–1 at
full scale in helium.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 209
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Total leakage test cycle times from 0.5 to instantaneously in standard engineering
2 s are obtainable. The flow compensation units and in the range selected. A self
network allows dynamic air flow check mode provides means for verifying
measurements without requiring a the integrity of the flow monitor both
stabilization period. electrically and pneumatically.
Digital Electronic Flow Vacuum Pumping
Meter for Monitoring Technique of Leak Testing
Leakage Rate Tests with Flow Measurements
Figure 33 shows a portable, digital If the test system can be safely evacuated,
electronic flow meter designed for fast, leakage can be measured directly by
accurate indication of leakage rates of means of flow meter with vacuum
pressurized components such as valves, pumping arrangement sketched in Fig. 34.
O-ring seals, pressure vessels, holding The system under test is evacuated
tanks, tank cars and processing vessels. It through an opened isolation valve
provides a broad range of flow rate connected to the vacuum pump inlet. The
measurements up to 2 × 102 Pa·m3·s–1 exhaust gases from the vacuum pump go
(2 × 103 std cm3·s–1) with a resolution of through a surge tank to the flow meter. A
one part in 2000. Repeatability is bypassing valve around the pump
indicated as ±0.2 percent of full scale and provides an alternative path between the
accuracy is ±1.0 percent of full scale isolation valve and the surge tank. Before
values. The instrument includes a broad performing the leak test, the vacuum
range, high accuracy solid state digital pumping system leak tightness is first
flow indicator, combined with an integral determined by closing the isolation valve
flow regulator and a digital pressure and measuring the rate of gas flow
indicator. The only additional equipment through the flow meter. If this flow is
required is a pressure source such as negligible, the isolation valve is then
instrument air or nitrogen for tests as opened and the flow meter readings are
pressures up to 400 kPa gage (0 to taken only after an equilibrium (constant
60 lbf·in.–2 gage) or 700 kPa gate (0 to flow rate) condition has been achieved.
100 lbf·in.–2 gage) and means for making
connections to the test unit. Once the The vacuum pressure in the system
system and object under test have been under test is adjusted by means of the
pressurized, operation is changed from bypass valve, which controls the backflow
the charge mode to the leakage test mode. of gas from, the exhaust port of the
The leakage rate indication is displayed vacuum pump to its inlet port. The lower
limit of vacuum pressure for which the
FIGURE 31. Arrangements for leak testing with thermopile air flow meter: (a) pneumatic bridge leakage testing arrangement
with thermopile flow meter arranged to measure difference in pressure between test object and an identical leaktight object
(reference volume); (b) inline leakage testing arrangement in which test part is pressurized, line valve is closed and leakage is
indicated by pressure drop in flow meter sensing element.
(a) Bridge arrangement
Regulator Valve Test part
Reference part
Valve Transducer
Valve
Air source
Flow indicator and alarm
(b) Inline arrangement
Regulator Valve Transducer
Test part
Air source
Flow indicator and alarm
210 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
vacuum pumping leak analysis technique verify the accuracy of the leakage test
is useful is in the range of 3 kPa (25 torr). results and instrumentation used in that
The lower limit of leak testing sensitivity test. It also verifies the validity of the
is about 0.1 Pa·m3·s–1 (1 std cm3·s–1) and is dewpoint and temperature sensor
mainly dependent on the availability of locations within the containment
suitable flow meters for the vacuum structure. Flow meters used in these large
pressure range used during the leak test. scale leakage rate tests include thermal
mass flow sensors, rotameters and
Sealed Volume Flow Meter integrating gas flow meters usually with
Leak Testing of Nuclear ranges of 25 to 700 Pa·m3·s–1 (0.5 to
Containment Systems 15 std ft3·min–1). These flow meters are
usually designed for the planned leak
Sealed volume leak testing techniques are testing conditions and they produce
also used on large volume systems such as readouts compensated to standard
nuclear containment systems. For this pressure and temperature conditions. The
application, this procedure is commonly accuracy of the flow meter must be
called a verification test. Its purpose is to
FIGURE 33. Portable digital electronic flow meter for
FIGURE 32. Air flow meter with orifice and differential monitoring leakage rates in pressurized systems.
pressure transducer: (a) photograph; (b) pneumatic circuit.
(a)
FIGURE 34. Arrangement for vacuum pumping technique of
leakage measurement with flow meter.
(b) Dial Quick
gage disconnect
Air
supply Pressure Isolation Bypass
transducer
System valve valve
under
test
Out In Test
item
Pressure Surge
regulator tank
Solenoid Pump Flow meter
valves
Quick
disconnect
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 211
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
commensurate with the accuracy of the section of its length. Sensor coils are
leakage rate test instrumentation and also wound around the flow tube on either
with the accuracy required in the side of the heater coil and are connected
containment leakage rate test results. in a bridge circuit. The zero flow, the
bridge circuit is balanced and the output
Procedures for Flow Meter signal is zero. With flow the sensor coils
Verification Test of Nuclear detect the resulting temperature
Containment Systems difference, which is proportional to mass
flow. The output electrical signal varies
The verification test is normally linearly with the gas flow rate. Signals can
performed as the last phase of a be used for measuring, recording or
containment test. It follows the test for controlling gas flow rates with valves and
the system measured leakage rate Qam an automatic controller. Sensors for
(usually given as a percentage of air mass specific gases such as air, nitrogen,
lost in 24 h). The flow meter is installed hydrogen, oxygen and helium are
in the system with a valve to isolate it
from the system under test. The FIGURE 35. Thermal mass flowmeter uses a true mass flow
verification test may be performed by sensor for measuring gas flow rates accurately: (a) sensor;
measuring either the out-leakage or the (b) principle of operation.
in-leakage that passes through the flow
meter. For either technique, a meter valve (a)
is placed downstream from the direction
of leakage flow through the flow meter, to (b) To power
minimize the pressure loss across the flow
meter. After opening the isolation valve supply
between the test system and the flow
meter, this metering valve is adjusted to Upstream Downstream Bypass sensor
produce a leakage flow through the flow temperature temperature tube
meter from (or into) the test system that sensor
is some required percentage (usually 75 to sensor
125 percent) of the allowable leakage rate
Qa for the system under test. Flow Bridge Amplifier
for ∆T
The leakage rate test of the 15 to 28 V detection 0 to 5 V direct
containment is then continued. After a direct current current and
period of 4 to 6 h with a minimum of ten 4 to 20 mA
sets of data, the combined leakage rate Qc
of the containment system and flow
meter and the leakage rate Q0 of the flow
meter are determined using the flow
meter readings. The difference between
these two leakage rates is Qc – Q0 = Q´am.
This difference Q´am in reading is then
compared to the leakage rate Qam
measured previously on the containment
test system alone, before the inflow or
outflow of air from the containment
through the flow meter. The two values
must agree with 25 percent of the
measured containment leakage rate Qam.
This is to say that Qam – Qam must be
equal to or smaller than 0.25 Qam.
True Thermal Mass Flow
Meters for Accurate Flow
Rate Measurements
The containment verification test just
described requires a true mass flow meter
that measures the mass of gas that passes
through it. Figure 35 shows a true mass
flow sensor element which does not
require temperature or pressure
compensation and provides ±1 percent of
full scale accuracy and linearity. The
sensor unit has a stainless steel flow tube.
A heater coil is wound around the center
212 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
available with ranges from 0 to 0.015 up
to 0 to 10 Pa·m3·s–1 (0 to 10 up to 0 to
5 × 103 std cm3·min–1). Repeatability of
indications is claimed as ±0.2 percent of
full scale. Output signals from the thermal
mass flow sensors of Fig. 35 can actuate
indicating meters or provide 0 to 5 V
direct current signals that can be
transmitted up to 300 m (1000 ft) to
recording instruments, digital indicators
or controllers. The electrical output signal
is linearly proportional to the mass flow
rate through the sensor.
Flow Meter Tests to Locate
Leaks in Gas Filled Electric
Power Cables
Electric utility companies have made use
of a U-tube manometer equipped with
appropriate valving to use as a flow meter
for locating gas leaks in gas pressurized
electric power cable sheaths. When the
manometer is installed in a segment of
the pressurized gas filled cable sheathing,
oil will rise in the glass tube of the
manometer, on the side closer to the leak.
In this test, the manometer measures the
pressure loss in the segment of cable
across which it is connected, when gas
flows toward the leak.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates 213
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
References
1. CRC Handbook of Chemistry and
Physics. Cleveland, OH: Chemical
Rubber Company (1964).
2. Fleshood, D.L. “Containment Leak
Rate Testing: Why the Mass-Plot
Analysis Method Is Preferred.” Power
Engineering. Barrington, IL: Technical
Publishing Company (February 1976):
p 56-59.
3. Lau, L.W. “Data Analysis during
Containment Leak Rate Test.” Power
Engineering. Barrington, IL: Technical
Publishing Company (February 1978):
p 46-49.
4. Kendall, M.G. and A. Stuart. The
Advanced Theory of Statistics, third
edition. Vol. 2. New York, NY: Hafner
Publishing Company: p 130-132.
5. Tietjen, G.L., R.H. Moore and R.J.
Beckman. “Testing for a Single Outlier
in Simple Linear Regression.”
Technometrics. Vol. 15, No. 4.
Alexandria, VA: American Statistical
Association (November 1973):
p 717-721.
6. ANSI/ANS-56.8-1981, Containment
System Leakage Testing Requirements,
Appendix C. La Grange Park, IL:
American Nuclear Society (1981).
7. Guthrie, A. Vacuum Technology. New
York, NY: John Wiley and Sons (1963).
Reprint, Malabar, FL: Krieger
Publishing (1990).
8. Steinherz, H.A. Handbook of High
Vacuum Engineering. New York, NY:
Reinhold Publishing Corporation
(1963).
214 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
6
CHAPTER
Leak Testing of Vacuum
Systems
Charles N. Sherlock, Willis, Texas
Carl A. Waterstrat, Varian Vacuum, Lexington, Massachusetts
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. The Nature of Vacuum
Definition of a Vacuum and computer chip manufacturing,
vacuum is used in magnetrons, cathode
The word vacuum is derived from the ray tubes, television picture tubes,
Greek word meaning empty. In practice, semiconductor devices, solar cells, plating
use is made of some type of vessel metals and plastics, thin film deposition,
(vacuum enclosure, chamber or container) lifting objects, plasma physics, cryogenics,
to contain a vacuum. When the enclosure metallurgical processing, electron beam
is closed to the surrounding atmosphere welding, brazing, distillation organic
and air or gas is removed by some chemistry, packaging, mass spectrometry,
pumping means, a vacuum is obtained. space simulation and leak detection.
Various degrees of vacuum can be Many other areas find application for
obtained, depending on how much air is vacuum equipment.
removed from the enclosure. Common
terms such as partial vacuum, rough Changes in Pressure Units
vacuum, high vacuum and ultrahigh Used for Vacuum
vacuum are used to describe degrees of Measurements
vacuum. A vacuum is any pressure below
the prevailing atmospheric pressure. The presently preferred SI unit for
Practically speaking, a vacuum such that pressure is the pascal (Pa). The standard
the containing vessel is empty, i.e., free of atmospheric pressure at sea level and 0 °C
all matter (molecules), is never obtained. (32 °F) is equal to 101.325 kPa. Earlier
If this were possible, the vacuum would units used for pressure in vacuum relate
be called a perfect or absolute vacuum. to atmospheric pressure indicated by the
height (nearly 760 mm) of the mercury
Applications of Vacuum barometer column at sea level and 0 °C
Environments (32 °F). The unit known as the torr was
defined as 1/760th of the pressure of the
Vacuum is used to reduce the interaction mercury column. The torr was named in
of gases or air with solids and to provide honor of an Italian physicist, Evangelista
control over electrons and ions by Torricelli (1608-1647), inventor of the
reducing the probability of collision with mercury barometer. The torr is almost
molecules of air. Vacuum pumps are used identical to the millimeter of mercury
by industry and laboratories to create a (mm Hg), because there are 759.96 torr in
vacuum environment for these a standard atmosphere. The difference
operations. Most gases react with solids to between the two units amounts to so little
cause effects such as oxidation, which it that torr and mm Hg have been used
may be necessary to avoid. In a vacuum interchangeably.
environment, the necessary operation
may be performed so that undesirable Variation of Atmospheric
effects are reduced or eliminated. For Pressure with Altitude
example, unless most of the air is
removed from an incandescent light bulb, The mercury barometer is a device for
oxygen in its atmosphere will react with measuring atmospheric pressure. As the
the hot tungsten filament, causing it to altitude increases, the pressure decreases
burn out prematurely. An electron tube because fewer gas molecules press on any
could not operate at atmospheric pressure. surface. A knowledge of how the pressure
Electron flow would be impeded by changes with altitude is very important in
collision with air molecules due to the connection with various space studies.
extremely small mean free path. In Table 1 shows the relationship between
addition, elements within the tube may pressure and altitude in the earth’s
react with the air. Other examples can be atmosphere.
cited where vacuum is necessary to
produce desired results that could be At an altitude of 50 km (27 mi) the
unattainable in any other way. pressure is about 0.1 percent of standard
atmospheric pressure or 100 Pa
Vacuum is required in many industries (0.015 lbf·in.–2). Air at this altitude
and products. In addition to light bulbs
216 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
contains one thousandth of the number ability of a gas to diffuse increases when
of molecules per unit volume in air at sea its pressure is reduced. Consider the
level. At 400 km (250 mi) altitude, the example of ammonia vapor being released
pressure is in the range of 1 µPa or 10–11 in a room. The reason that it is not
parts of sea level pressure. Table 2 gives a detected immediately at the other end of
relative measure of gas characteristics at sea the room is that the path each ammonia
level and at 1 nPa (10 ptorr). Compared molecule takes is restricted by the air
with the number of molecules in a cubic molecules with which it collides. It is only
centimeter at atmospheric pressure, it is after many billions of collisions with air
seen that there are one hundred molecules that the ammonia molecules
thousandth of one millionth as many finally make their way across the room. If
molecules at 10–6 Pa (1.5 × 10–10 lbf·in.–2). the room were pumped down a high
However, a tremendous number of vacuum, there would be many fewer air
molecules (3 × 108) still remain in a cubic molecules and far fewer collisions to
centimeter at a pressure of 1 µPa (10 ntorr). impede the path of the ammonia
Pressures around 1 µPa (10 ntorr) are not molecules. Thus, an ammonia molecule in
uncommon in good vacuum systems. a high vacuum takes less time to complete
its trip across a given distance than in
Diffusion and Adsorption gases at higher pressures.
of Gases in Vacuum
Systems Only those molecules that are in
motion within a vacuum chamber create a
The kinetic molecular theory of gases and pressure through collisions with its walls.
the ideal gas laws (Boyle’s, Charles’, A molecule that is adsorbed to the wall
Dalton’s and the general gas law), are surface is stationary and does not produce
applicable to vacuums. In vacuum, fewer collisions. Therefore, adsorbed gas
molecules are dealt with, but their basic molecules do not contribute to the total
behavior is predictable by the molecular pressure. However, molecules adsorbed on
theory of gases and does not change. The surfaces can be returned to the gas phase
by thermal agitation produced by the
TABLE 1. Change in atmospheric pressure with altitude. application of heat. Thus, outgassing
effects can contribute to pressure in an
____A__lt_it_u_d__e____ _________P_r_e_s_s_u_r_e_________ evacuated system, because a molecule can
undergo repeated collisions and exert
km (mi) kPa (atm) Remarks pressure only when it is in the gaseous
state.
0 (0.6) 101.325 (1.00) international
(1.2) standard Mean Free Path of Gases
1 (3.1) 89.90 (0.887) in Vacuum Systems
2 (6.2) 79.50 (0.785) jetliner
5 54.00 (0.533) altitude At normal atmospheric pressure, gas
10 (12.4) 26.50 (0.262) molecules make many collisions with
(31.1) low orbit each other. The average distance that a
20 (62) 5.53 (0.055) molecule travels before colliding with
50 (124) 7.98 × 10–2 (7.9 × 10–4) another molecule is known as the mean
100 (311) 3.2 × 10–5 (3.2 × 10–7) free path. The mean free path of two
200 (621) 8.5 × 10–8 (8.39 × 10–10) different gases at the same pressure will
500 3.0 × 10–10 (3 × 10–12) not be the same; this is because the mean
1000 7.5 × 10–12 (7.4 × 10–14) free path depends on the molecular size,
which varies from one gas to another. In
spite of this fact, it is still possible to give
a useful relationship between mean free
path and pressure. The approximate
values of mean free paths for air and
TABLE 2. Comparison of atmospheric properties at sea level and at high altitude.
Condition At Sea Level At 400 km (250 mi) Altitude
Pressure 101.325 kPa (760 torr) 1 µPa (10 ntorr)
Number of molecules in 1 cm3 (0.06 in.3) 2.7 × 1019 3 × 108
Mean free path
Time to form a monolayer of adsorbed gas on a clean surface 93 nm (3.7 × 10–6 in.) 9.3 km (5.8 mi)
Average speed of nitrogen molecule at room temperature 20 °C (68 °F) >10 ns 120 s
1600 km·h–1 (1000 mi·h–1) 1600 km·h–1 (1000 mi·h–1)
Leak Testing of Vacuum Systems 217
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
other gases are given as a function of gas flow. The motion of a particular molecule
pressure in Eq. 1: is entirely random and unpredictable; it is
as likely to move in one direction as in
0.0095 any other direction. To a molecule, tube
(1) λ = wall appear very rough and irregular. The
direction of molecule rebound after
P impact with the tube wall thus tends to
be independent of the direction of
where λ is mean free path (meter) and P is incidence. (This is an over simplified
gas absolute pressure (pascal). description.) Figure 1 is a sketch of
particle motions during molecular flow of
Effects of Molecular Friction and gases through a tube. Note that not all of
Gas Viscosity in Viscous Flow the molecules entering at the left exit at
the right. The gas flow will continue as a
As shown by Eq. 1, the mean free path net movement to the right only as long as
length varies inversely with absolute there is some driving force causing
gaseous pressure. The concept of mean movement from left to right. As gas
free path is useful in describing vacuum concentration gradient is such a force. A
ranges. The mean free path at pressure differential is another force that
atmospheric pressure is very short (see can control the direction of net flow of a
Table 2), due to the large molecular gas. Both can contribute to flow of a
density. Therefore, collisions occur much tracer gas through a leak.
more frequently between gas molecules
than between molecules and the walls of Specifying Gas Flow Rates
the container. Thus, the gas acts much
like a fluid. Under a pressure differential The flow rate of liquids is expressed
this gaseous fluid moves as a unit and is simply as so many volume units per unit
considered to flow. The molecules, while time, such as liters per second. When,
drifting slowly in the direction of flow, however, the flow rate of gases is
move rapidly along random paths. Any considered, it is necessary to know not
resistance to this flow is due to the only the volume of a gas but its pressure
viscous properties of the gas. The term and temperature as well. A cubic meter
viscous refers to molecular friction and is volume of gas at 100 kPa (15 lbf·in.–2)
used to describe the flow characteristics of pressure and a temperature of 20 °C
a fluid. Water, for example, is less viscous (68 °F) will contain ten times as many
than syrup because it flows or pours more molecules as a cubic meter volume of gas
easily. The cross sectional dimension of at 10 kPa (1.5 lbf·in.–2) and 20 °C (68 °F).
the container or tube through which the Only a complete statement of volume,
gas flows is important because it displacement rate, gas pressure and
determines the velocity of the molecules temperature can accurately describe the
within the flowing gas. The viscous total quantity of gas that flows per unit of
properties of the gas are functions of the time. In both liquids and gases, it is mass
gas viscosity and the gas velocity. When flow that is of interest. For liquids of
viscous properties control gas flow rates, constant density, the mass rate of flow is
the situation is termed viscous flow. directly proportional to volume flow rate.
With gases, density varies both with
Effects of Mean Free Path and temperature and with pressure. Thus, for a
Flow Cross Section on Molecular given gas, volume displacement rate,
Flow of Gas pressure and temperature must be known
to define the mass flow rate.
As the pressure of the gas within a system
is reduced, the mean free path of the FIGURE 1. Molecular motion along a tube, with particle mean
molecules increases and the flow free path far larger than tube diameter.
characteristics change gradually. As the
mean free path becomes comparable to 3 1
the cross sectional dimensions of the 5
tube, collisions occur less frequently 2
between molecules and the apparent 4 3
viscosity of the gas decreases. Under these
conditions, the event that is most likely 2
to affect the direction of the molecules; 5
travel is a molecular collision with the 41
tube wall. As the pressure is further
reduced, the mean free path becomes
greater than the tube’s cross sectional
dimensions. The diameter of the tube
alone then determines the resistance to
flow; this situation is called molecular
218 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The Concepts of Gas By combining Eqs. 2 and 3, the product of
Quantity and Pumping pumping speed S and gas pressure P can
Speed be equated to throughput by Eq. 4:
From the gas laws, it is known that the (4) Q = S × P
product PV of pressure P and volume V is
proportional to the number of molecules Equation 4 is the universal relationship
in a sample of gas. In static systems, the on which vacuum pumping throughput
PV product is constant at a given calculations are based. As an example of
temperature. This product PV is known as its use, suppose the gas in the pipe
the quantity of gas. Common units of gas between Sections 1 and 2 of Fig. 2 passes
quantity include torr liter (torr-L); the Section 1 in 1 s and this volume V is
atmospheric cubic centimeter (cm3 of 100 L (0.1 m3) and pressure P at Section 1
volume at standard sea level atmospheric is 10–4 Pa and displaced volume V =
pressure or std cm3); and the bar liter 0.1 m3, divided by the time t = 1 s:
(bar-L). The preferred SI unit of gas
quantity is the pascal cubic meter (Pa·m3). Q = S × P = PV
t
In steady flow, the same quantity of
gas (number of molecules) that enters one = 10−4 × 0.1 = 10−5
end of a tube must leave at the other end,
even though there may be different Comparison of Gas Flow with
volumes of gas entering and leaving per Liquid Flow
unit time. If the PV product is used as a
measure of the amount of gas flowing Before attempting a more thorough
through a tube, computation may be discussion of gas flow, it may be helpful
done with a minimum of complication. to compare gas flow with water flow. To
get any fluid to flow within a pipe, a
The volumetric pumping speed S is the pressure differential must be established
time rate of volume displacement, as between the two sections across which
given by Eq. 2: the fluid is to flow. (Gravitational effects
are neglected in this introductory
(2) S = V discussion.) The fluid would then flow
t from the high pressure region P1 to the
low pressure region P2. Consider a closed
Typical units of pumping speed S system of pipes through which water is
would be cubic meter per minute circulated as in an automobile. The water
(m3·min–1), cubic meter per second pump creates the pressure differential
(m3·s–1), liter per second (L·s–1) and cubic necessary for water to flow. Across each
foot per second (ft3·s–1). component (radiator, engine block,
thermostat, different sizes of piping) in
Concepts of Throughput the system, the pressure drops. The sum
and Leakage Rate of all these pressure drops equals the
pressure differential across the water
In vacuum practice, the preferred pump. The magnitude of the pressure
description of the rate of flow of gas is drop across each component of the
commonly called throughput. system depends on its physical geometry.
Throughput is the quantity of gas or a Clearly, a smaller diameter pipe will result
measure of the total number of molecules in decreased flow for the same size pump.
at a specified temperature, passing an similarly, increasing the length of the pipe
open section of the vacuum system per will reduce the flow, whereas decreasing
unit time. Leakage rate is a similar the length of the pipe will increase the
measure of the total number of molecules flow. Shorter lengths and larger diameters
at a specified temperature passing through reduce the resistance to flow through the
a leak per unit time. Q is the symbol pipe.
commonly used for gas throughput per
unit time, in pascal cubic centimeter per Analogy of Gas Flow to Electric
second: Current through Resistance
(3) Q = PV The analogy may be carried further by
t comparing the gas flow system to an
electrical circuit. Given an electrical
circuit with a battery and a resistor in
series with it, the battery may be
considered to be the pump and the
resistor the pipe. Increasing or decreasing
the resistance decreases or increases the
current flow (analogous to gas flow),
Leak Testing of Vacuum Systems 219
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
respectively. If the circuit consists of a In an electrical circuit, the voltage drop
series of resistors and a battery, the sum of across a resistor is the product of the
the voltage drops across each of the current and resistance. In a vacuum
resistors (pressure drops) is equal to the circuit, the pressure differential across a
total voltage generated by the battery pipe is the product of throughput (gas
(pressure differential created by the flow) Q and resistance R. Equation 5 states
pump). The voltage drop across each this relation mathematically for the
resistor will depend on the magnitude of pressure differential ∆P:
resistance of that component. The larger
the resistance, the larger the voltage drop (5) ∆ P = P1 – P2 = Q×R = Q
(see Fig. 3). C
Gas Conductance and Its Electrical Because R is equal to 1/C, Eq. 5 may be
Analogy written in the form of Eq. 6 for gas
conductance C:
In vacuum, one speaks not of the
resistance a tube or component offers to (6) C = Q
gas flow, but instead uses the reciprocal ∆P
term conductance. Conductance is a
measure of the ability of a vacuum Equation 6 is the defining equation for
component to permit gas flow or not to gas conductance: the ratio of throughput
impede it. Consequently, the greater the Q to pressure differential ∆P across the
resistance, the smaller the conductance conductance.
and vice versa. Figure 3 shows the
electrical analogy of a tube in a vacuum Gas Conductance with
system. The battery is analogous to the Sequential Tubes of
vacuum pump, current is analogous to gas Passages
flow and the resistor is analogous to pipe.
FIGURE 2. Rate of flow of a gas Q through tube with applied If two different diameter pipes with
different conductance values are
pressure differential ∆P = P1 – P2 (P1>P2). (See analogous connected in series as in Fig. 4a, the total
electric circuit of Fig. 3.) conductance of the connection between
extreme ends decreases (resistance
Section 2 Section 1 increases). From Eq. 6, the conductance of
P1 the pipe between Sections 1 and 3 may be
P2 expressed as in Eq. 7:
To vacuum
Tube conductance = C (7) C13 = Q
pump Tube resistance = R P1 − P3
From vacuum
chamber
Gas flow rate ( ) ( )(8) P1 − P3 = P1 − P2 + P2 − P3
Gas flow rate Q = (P1 – P2)C (9) P1 − P2 = Q
C12
or
Pressure differential ∆P = P1 – P2 = QR = Q/C Q
C 23
P2 − P3 =
FIGURE 3. Electrical analogy of vacuum pumping pipe Now, by combining Eqs. 7, 8 and 9, the
conductance system of Fig. 2. G is the electrical relationship for C13 becomes:
conductance, the reciprocal of electrical resistance R, so that
G = 1/R. ER = Eb – EL = E1 – E2 = IR = I/G. (10) C13 = Q
Q +Q
R = 1/G C12 C23
or, in its reciprocal form:
– i 1 Q +Q
Eb EL Load = C12 C 23
(11)
+ C13 Q
= 1+1
C12 C23
220 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
In its general form, Eq. 11 may be written division is not equal but depends on the
as Eq. 12: conductance of each component. From
Eq. 6, the gas load in each parallel pipe
(12) I = 1 + 1 may be written in the form of Eq. 14:
CT C1 C2
+ 1 +…+ 1 (14) QA = CA × ∆ P
C3 Cn and
In Eq. 12, the subscript T denotes the Qb = Cb × ∆P
total conductance of a number of
conductances C1, C2, C3 ... Cn connected The total conductance between points 1
in series. and 2 is
In the case of only two conductances (15) C12 = Qa + Q b
connected in series (Fig. 4b), Eq. 12 ∆P
should be written in the form of Eq. 13:
C1 × C2 Substituting from Eq. 14, Eq. 15 gives:
C1 + C2
(13) CT = (16) C12 = Ca ∆P + Cb ∆P
∆P
This case is analogous to the special case Simplifying, Eq. 16 becomes:
of two electrical resistors connected in
parallel. (17) C12 = Ca + C b
Gas Conductance for Pipes In its general form, the total conductance
or Tubes Connected in for a number of pipes connected in
Parallel parallel is equal to the sum of the
individual conductances, as given by
Figure 5 shows two lengths of pipe Eq. 18:
connected in parallel. In this connection,
the total gas load (throughput) flowing (18) CT = C1 + C2 + C3 + … + Cn
from the vacuum chamber divides
between the two pipes as shown. The FIGURE 5. Electrical analogy of two gas conductances in
parallel: (a) connection of two parallel gas conductances;
FIGURE 4. Electrical relationship of two conductances in (b) electrical circuit analogous to two gas conductances in
series: (a) pipe conductances in series; (b) connection of two parallel.
electrical resistances.
(a) P1
Ca
P2
(a) P2 P1 Qa Pipe 1
Cb
To P3 C23 C12 To turbomolecular
turbomolecular or diffusion pump Qb Pipe 2
or diffusion QT = Qa + Qb
pump
Q Q Vacuum From
Pipe 2 Pipe 1 chamber vacuum
chamber
—1 = —1 + —1
C13 C12 C23
(b) R2 R1 (b) R1
I EB – I1
(pump) R2
EB – Load
(pump) + + I2
EL (vacuum Load
chamber) EL (vacuum
EB
chamber)
Leak Testing of Vacuum Systems 221
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Pumping Speed throughput Q is the product of the speed
S and pressure P where each is measured
In operating a vacuum system, there is an at the same point, such as at the pump or
interest in how fast gases are removed chamber. Throughput at the pump is
from the system. The rate of removal of therefore expressed as Eq. 21:
gases is measured by pumping speed S.
From Eq. 4, pumping speed is defined as (21) Q p = S p Pp
the ratio of the throughput Q to the
pressure P at the point in the system. and pressure as:
Mathematically, this relation is given by
Eq. 19 (m3·s–1): Pp = Q
Sp
(19) S = Q
P
If the inlet to a vacuum pump were At the chamber being evacuated,
connected directly to a vacuum vessel, throughput is expressed as:
then the pumping speed at the vessel
would be the same as that at the pump (22) Qc = Sc Pc
inlet. Because it is physically impossible to and pressure as:
join the pump and vessel without
introducing a connector the pumping Pc = Q
speed at the vessel will be lower than that Sc
at the pump. Pumping speed loss depends
on the magnitude of the conductance Substituting Eqs. 21 and 22 into Eq. 20
that causes a loss in pressure or creates results in the relation of Eq. 23:
differential pressure between pumps and
vessel. Figure 6 is used to help establish a Q –Q =Q
relationship between the net pumping (23) Sc Sp C
speed at the vacuum chamber, pumping
speed at the port of a vacuum pump and Rearranging terms:
the conductance between them. Although
the connection is shown as a pipe in Q =Q +Q
Fig. 6, it could be a combination of any (24) Sp C
number of vacuum components, each
contributing a value of conductance. The Sc
flow of gas is from the chamber to the
pump. From Eq. 5, the pressure drop is and multiplying by 1/Q:
given by Eq. 20:
(20) ∆ P = Pc − Pp = Q (25) 1 = 1 + 1
C Sc S p C
In Eq. 20, the subscripts c and p refer to In the general case, the net speed Sn at
the chamber and pump, respectively. The any point in a vacuum system is related
FIGURE 6. Net pumping speed relationship applicable to to the pump speed Sp and the total
conductance Ct between that system point
conductance C between vacuum pump and chamber being and the vacuum pump by Eq. 26:
evacuated. Pressure at vacuum chamber (inlet to (26) 1 = 1 + 1
Sn S p Ct
conductance C ) is Pc , and pumping speed Sc = Q/Pc.
Pressure at vacuum pump (outlet of conductance C) is Pp Analysis of Eq. 25 shows that, except for
and pumping speed is Sp = Q/Pp. Because vacuum pump the case of an infinite conductance (zero
pressure Pp is lower than chamber pressure Pc , whereas Q is resistance), the net speed will always be
the same at each end of conductance C, the pumping speed less than the pump speed. How much less
depends on the value of the tube
is different at inlet and outlet of C. conductance.
Diffusion C
pump Vacuum
chamber
S—n = Qn Q Qc = —Sc
Pn Pc
Sp = —Q Sc = —Q
Pp Pc
222 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 2. Principles of Operation of Vacuum
Systems and Components
atomic mass unit (u); and T is the
Introduction to Vacuum absolute temperature of gas within the
Pumping
evacuated container (kelvin). A similar
To attain vacuum in a container, some
means of pumping must be used. A pump expression can be given for a mass of gas
cannot reach into the system and extract
molecules, but must simply wait for (kilogram) striking a unit area (square
molecules to wander through the natural
exit in the container and into the pump meter) of the evacuated container wall per
for isolation and discharge. When the
pressure in the vacuum system becomes unit time (second):
so low that there is no longer a net
movement of molecules into the pump, ( )R’ = 43.8 × 10−4 P
the base pressure of the system has been
reached under those conditions. How low MT
this ultimate pressure is will be
determined by conditions such as (1) the where R is rate of gas mass impact with
leak tightness of the vacuum system, wall (kg·m–2.s–2); P is absolute pressure
(2) the nature and condition of materials within evacuated chamber (pascal); M is
within the vacuum system that might molecular weight of gaseous particles, in
cause outgassing and (3) the operating unified atomic mass unit (u); and T is
characteristics of the pumps in absolute temperature of gas within
combination with the specific vacuum evacuated container (kelvin).
system.
If a hole of unit area were cut through
Molecular Conditions the container wall, those gas molecules
Limiting Rates of Pumping that would have collided with the
of Vacuum Systems container wall in the area of the hole and
rebounded within the container will now
To evacuate a closed system initially at pass through the hole at the rate given by
atmospheric pressure, numerous gaseous Eq. 27a. If these escaping gas molecules
molecules must be removed from within are now prevented from reentering the
the closed system. The fewer the gas container through that hole, the net effect
molecules remaining, the lower the would be that of reducing the number of
absolute pressure of gas within the molecules within the container and thus
system. However, the common concept reducing the internal gas pressure. This is
that vacuum pumps draw out the air like the basic concept of vacuum pumping,
a vacuum cleaner is wrong. Molecules in namely, to provide a natural exit for gas
the gaseous phase are in constant motion molecules and to isolate the escaping gas
and collide with each other and with the molecules so that they cannot reenter the
walls of the container. A certain number R container being evacuated. The vacuum
of molecules strike each unit area of the pump cannot extract gas molecules from
container wall per unit time. The number within the evacuated container; it merely
of gaseous molecules striking a unit area aids those molecules that pass through
(square meter) of the container wall per the hole in the wall to naturally escape
unit time (second) is given by Eq. 27: being reinjected into the vacuum
container through that same hole. It is
impossible to pump any gas out of an
evacuated container at any rate faster
than that at which internal gas molecules
strike the hole area by their random
kinetic motions.
( )(27) R = 2.63 × 1024 P Example of Limitations on Vacuum
Pumping of Gaseous Nitrogen
MT
For gaseous nitrogen with a molecular
where R is rate of molecular impact with weight M = 28 at room temperature,
the wall in molecules per square meter per 298 K (25 °C or 77 °F) and atmospheric
second; P is absolute pressure in evacuated pressure, 101 kPa (1 atm), Eq. 27a
chamber (pascal); M is the molecular indicates that the number of molecules
weight of gaseous particles, in unified striking each square meter of container
wall during each second would be
calculated as:
Leak Testing of Vacuum Systems 223
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
R= ( )2.63 × 1024 × 101 000 Operation of Mechanical
Pumps for Vacuum
28 × 298 Systems
= 2.92 × 1027 The mechanical pump is an essential
component used in vacuum systems to
Similarly, if the mass (kilogram) of evacuate a chamber from atmospheric
nitrogen gas striking the unit area (square pressure to about 0.1 Pa (10–3 torr)
meter) per second were to be determined, absolute pressure. Of the various types of
Eq. 27b indicates that this mass would be: mechanical pumps, the rotary oil sealed
vacuum pump shown in Fig. 7 is most
( )R’ = 43.8 × 10−4 × 101 000 common. The pump consists of a
stationary housing, an eccentrically
× 28 mounted rotor with two spring loaded
298 vanes, an inlet port and a discharge port.
Air enters the pump from the vacuum
= 136 chamber through the inlet port. This air is
trapped, compressed and ejected into the
The volume flow rate corresponding to atmosphere through the discharge port by
means of the rotor arrangement. Sealing
the above mass flow rate would be equal of the eccentric rotor vacuum pump is
to 136 × 0.8714 = 118 m3·s–1·m–2 done by an oil film between the two
(7.08 × 103 ft3·min–1·ft2). sliding spring loaded vanes that make
contact between the rotor and the
For the case of molecular nitrogen (N2) housing. Oil is used as the pump sealant.
at 298 K (77 °F) and atmospheric pressure, Close tolerances must be maintained to
the conversion factor is 0.8714 m3·kg–1. prevent leaks and by passing of gases.
Consequently, care must be taken to
Thus, for gaseous nitrogen molecules prevent solid particles from entering the
pump. Each rotation of the rotor
(each of which contains two nitrogen discharges two volumes; each volume is a
certain percentage of the volume to be
atoms), the volumetric rate of exit of gas evacuated. This would indicate that even
a perfect pump could never evacuate to a
through a hole in the evacuated container vacuum linearly but could only approach
would be 118 m3·s–1·m2 or 11.8 L·s–1·cm2 this condition as an exponential function
(7.08 × 103 ft3·min–1·ft2). This is the of pumping time.
maximum rate at which nitrogen can be
FIGURE 7. Rotary mechanical vacuum pump with eccentric
pumped from a container whose internal rotor and spring loaded vanes. Pump oil provides a sealing
film at points of vane contact with stator housing.
pressure was 101 kPa (1 atm). As system
pressure decreases, the molecular or mass
rate of pumping drops proportionally.
Conditions Limiting Rate Outlet Inlet
of Pressure Reduction by Oil
Pumping Rotor
Pumping times greater than expected for Vane
reduction of pressure to desired levels can
result from system contamination or Spring
system leaks. System contamination can
be caused by processing of so-called dirty Housing
work materials or by allowing excessive
time without thorough cleaning of the
vacuum equipment. Contamination of
this type results in many layers of various
compounds, organic or otherwise, which
build up on interior surfaces. The
contaminated surfaces then outgas at such
rates that the pump capacity may be
unable to reduce pressure to desired levels
within acceptable pumping times. Water
vapor adsorbed to chamber walls is a
common contaminant. Dirty walls are
subject to more severe water adsorption. It
is also possible for mechanical pump oil
to become contaminated, which alone
can cause poor pumping characteristics.
If pumping is slowed by system leaks,
thorough mass spectrometer leak
detection tests inspection should be
performed and leaks repaired.
224 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
At low chamber pressures, air may leak closed because there is usually very little
back into the evacuated volume. This can water vapor present. The minimum
be minimized by putting two pumps in pressure obtainable is also slightly higher
series so that the discharge from the first with the ballast valve open. (Because of
pump chamber is not directly to the higher pressure in the chamber during
atmosphere but to some intermediate compression with the ballast valve open,
pressure maintained by the second (or there is more leakage back into the
backing) pump. vacuum system.) Some pump
manufacturers recommend operating the
Pump Oil Used in Rotary Vacuum pumps with the ballast valve open once
Pumps each week for about 20 min to drive out
any water vapor that may have
The operating fluid in any type of pump accumulated in the pump oil.
is called the pump fluid or pump oil.
With rotary pumps, normally a good Ultimate Pressure
quality light petroleum oil, with the high Attainable in Rotary Pump
vapor pressure factions removed, is used Vacuum Systems
to provide pump sealing and lubrication
between the rotor vanes and stator The limiting absolute pressure approached
housing. The oil for lubricating and in a vacuum system, after sufficient
sealing is contained in an oil reservoir. pumping time establishes that further
The arrangement of the reservoir differs reductions in pressure will be negligible, is
from manufacturer to manufacturer. In called the ultimate pressure. The range of
some small pumps, the pump chamber is ultimate pressures of commercial rotary
actually immersed in the reservoir, vacuum pumps extends from about 3 mPa
whereas for the larger pumps the reservoir to 1 kPa (20 µtorr to 5 torr). The low
is usually separated from the pump pressure of 3 mPa is reached only under
chamber, often being mounted above the the most ideal conditions. The ultimate
pump itself. pressure will be determined by:
Prevention of Condensate 1. outgassing of the pump,
Contamination of Pump 2. the seal between rotor and stator,
Oils 3. contamination of pump oil and
4. the vapor pressure of the oil used.
Contamination of pump oil is one of the
main difficulties with rotary pumps. As A high vapor pressure pump oil will
the gases and vapors are compressed, the evaporate at a greater rate, which will
vapors will tend to condense and create gas loads that saturate the pump
contaminate the oil. Degassing of vapors and limit the ultimate pressure attainable.
from pump oil can limit the ultimate
vacuum attainable. Pumps are available A disadvantage of any oil sealed and
with a gas ballast valve incorporated, lubricated pump is the backstreaming of
which minimizes the condensation of oil vapors from the pump inlet when inlet
vapors in the pump oil. The gas ballast pressures drop below or approach 70 Pa
valve is a small valve that can be opened (0.1 torr). This has become a major
manually to admit a controlled amount of concern to many industries, such as
air to the pump cylinder during part of semiconductor producers, for whom
the compression cycle. This will dilute the backstreaming causes contamination of
vapors to the point where they do not their products with oil vapors. As a result,
condense during compression. The violent several new pump designs classified as dry
agitation of the oil by the additional air or relatively free of this problem have
rushing through the pump causes been available since the early 1980s. Two
reevaporation and exhaust of water that of these are called scroll pumps and hook
may have been pumped from the vacuum and claw pumps.
system in vapor form and condensed in
the pump oil. To effect the removal of Rotary Dry Mechanical
moisture when the surrounding air is Pumps
saturated with moisture, connect a dry
nitrogen gas supply to the gas ballast. Be Unlike the rotary vane pump, which
careful to select a nitrogen flow rate and requires a low vapor pressure oil to
pressure that will not apply overpressure lubricate and seal the internal surfaces,
to the casing of the pump. The extent of two commonly used pumps are designed
use of the ballast valve is determined by with very small clearances between the
the amount of such vapors handled by moving and fixed surfaces and no need
the pump. In normal high vacuum for oil. As a result, the contamination
service, the ballast valve is usually kept caused by vapors entering the evacuated
space at low pump inlet pressures is
Leak Testing of Vacuum Systems 225
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
eliminated. A slight disadvantage is that pumping, gas is drawn in one side of the
these pumps cannot quite reach the same claws and compressed on the other as the
low pressure as the lubricated pumps. claws rotate. During rotation, the right
Ultimate pressures for dry pumps is in the claw opens the intake slot, allowing gas to
range of 2 to 3 mPa (20 to 30 µtorr). be drawn into the chamber.
Simultaneously, the left claw opens the
Rotary Scroll Mechanical Pumps exhaust slot letting compressed gas
escape. On completion of the
A scroll is a free standing involute spiral compression cycle, the claws pass through
with a solid base on one side. A scroll set, a neutral position and cycle begins again.
the fundamental element of a scroll
vacuum pump, is made up of two Pumping Speeds of Rotary
identical right and left hand involutes. Mechanical Vacuum Pumps
When assembled, one scroll is indexed
180 degrees with respect to the other, to Apart from the ultimate absolute pressure
allow the scrolls to mesh (Fig. 8). that can be achieved by an particular
pump, there is an interest in how fast the
In operation, one scroll is fixed and the pump can reduce the pressure in a
other is attached to an eccentric, driven vacuum system to the operating level.
by an electric motor. The pump inlet is at Manufacturers normally specify the
the periphery of the scrolls. As the pumping speeds of their mechanical
moving scroll orbits (but does not rotate) pumps at atmospheric pressure. In
about the fixed scroll, the entering gas is general, rotary pumps start pumping at
trapped in two diametrically opposed, atmospheric pressure and, as the pressure
crescent shaped pockets bounded by the is reduced, the pump becomes less
involutes and base plates of both scrolls. efficient. It then is pumping the same
The pockets shrink as they follow the volume, but at lower pressure. Eventually,
involute spiral toward the center, the pumping speed becomes zero at the
compressing the gas. The compressed gas ultimate minimum pressure.
exhausts to atmosphere through the
discharge port at the center of the fixed Figure 9 is a plot of pressure as a
scroll. function of pumping speed for a
400 L·min–1 (15 ft3·min–1) mechanical
Rotary Claw Mechanical Pumps pump. It is seen that at atmospheric
pressure, the pump is rated at
The hook-and-claw mechanism consists of 400 L·min–1; at 0.1 Pa (1 mtorr), the
several inline stages. The claw devices do pumping speed is 200 L·min–1 at 0.01 Pa
not make contact with each other or with (0.1 mtorr), the pumping speed is
the chamber walls, obviating oils. Each 16.7 L·s–1 (35 ft3·min–1). The pump speed
rotation of a claw pair consists of three reduces to zero at 10–3 Pa (10 µtorr), the
cycles: a start cycle, compression cycle ultimate pressure attained by this pump.
and a finish cycle. The two claws, which At this point the gas handling capacity
divide the pump chamber, turn in has been saturated by the gas load from
opposite directions and, in so doing, open the pump, thereby reducing its effective
and close the intake and exhaust slots pumping speed to zero.
through which the gases pass. During
Blower Pump or Booster Pump
FIGURE 8. Position of orbiting scroll shown before
compression cycle. The blower or booster pump (Fig. 10) is a
high throughput, low compression pump.
Orbiting scroll This pump is usually used on systems
where a large volume of gas must be
Pocket of gas pumped. It is also used with a mechanical
isolated pump to serve as the forepump for large
diffusion pumps, turbomolecular pumps
Gas inlets or even other blower pumps.
shown
closed The pump consists of two figure eight
shaped rotors or lobes mounted axially on
Gas inlets parallel shafts, as shown in the drawing
shown closed below. These rotors are synchronized by
gears to prevent physical contact and
damage and rotate in opposite directions.
This rapidly displaces gas from the inlet to
the outlet.
Exhausts at Fixed scroll
center
opening
226 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
How the Pump Works Vacuum System Operation
These rotors are designed so that, while Operating procedure consists of turning
spinning, they approach each other and the mechanical pump on, then the blower
the housing within several thousandths of (Fig. 11). Usually the mechanical pump
an inch. (See Fig. 10.) Rotor speeds vary has lowered the pressure sufficiently for
from 40 to 60 s–1 (2500 to 3500 rotations the blower to begin pumping by the time
per minute). Because of the high speeds the blower has reached operating speed. A
and close tolerances of the rotating lobes, bypass valve around the blower is
booster pumps are usually not started sometimes used for high pressure
until roughing pressures of about 1.3 kPa roughing.
(10 torr) have been reached. The typical
blower windmills at atmospheric pressure, Blowers are commonly used where
producing much heat and very little large volumes of gas need to be pumped.
pumping action. They are used when the lowest pressure
needed is 10–2 to 10–3 Pa (75 to 7.5 µtorr).
Blower or booster pumps are most They also are used to help the mechanical
useful in the 0.1 to 0.01 Pa (1.0 to forepump or backing pump maintain a
0.1 mtorr) pressure range. They are always low pressure and help reduce the
backed by a mechanical pump as a result. possibility of oil backstreaming.
Operating at high pressures will cause
heating and expansion of the lobes. This Turbomolecular Vacuum
can result in damage to the pump. No oil Pumps
is used to seal the gap between stator and
rotor. Oil is used in the forevacuum The turbomolecular pump serves as an
section of the pump to lubricate the gears alternative to the diffusion pump and
and bearings located there. must also be backed by a forepump. Its
FIGURE 9. Mechanical pump speed as a function of gas pressure for a pump rated at 6.7 L·s–1
(14 ft3·min–1) at atmospheric pressure.
Pump speed, L·s–1 (ft3·min–1) 7 (14.8) Atmospheric pressure
6 (12.7)
5 (10.6)
4 (8.5)
3 (6.4)
2 (4.2)
1 (2.1)
10–3 10–2 10–1 100 101 102 103 104 105
(10–7) (10–6) (10–5) (10–4) (10–3) (10–2) (10–1) (1) (10)
Pressure, Pa (lbf·in.–2 × 1.45)
FIGURE 10. Blower pump operation: (a) at beginning of cycle; (b) after eighth of cycle; (c) after fourth of cycle; (d) after three
eighths of cycle.
(a) (b) (c) (d)
Inlet Outlet to
forepump
Leak Testing of Vacuum Systems 227
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
principle advantage over the diffusion dimensions of the grooves at the inlet
pump is that it provides an essentially port must be such that the molecules
vapor free vacuum without baffles or cold have a good chance of hitting the walls of
traps. Thus, for a system where the back the groove or the blades without making
streaming of vapor from a diffusion pump numerous collisions with other gas
is undesirable or intolerable, a molecules (see Fig. 12). As the gas is
turbomolecular pump could be used. Its compressed while passing through
main disadvantage is that it has high successive stages of the turbine, it is
speed rotating parts whereas the diffusion necessary to decrease the dimensions of
pump has not moving parts. It also the air passages to keep them comparable
requires air gap tolerance on the order of with the mean free path of the molecules.
2 to 5 µm (8 × 10–5 to 2 × 10–4 in.) The system must already be evacuated by
between the high speed rotor and grooves a forepump before a turbomolecular
in the stator. As with a diffusion pump, a pump can start pumping. It can achieve
molecular pump cannot operate at pressures as low as 1.0 to 0.1 µPa (10 to
pressures above 13 to 1.3 Pa (100 to 1.0 ntorr). Pumping speeds for air vary
10 mtorr) and must be backed by a from about 70 to 9000 L·s–1 (1.5 × 102 to
mechanical forepump. 1.9 × 104 ft3·min–1), depending on the size
of turbomolecular pump selected.
A turbomolecular pump (see Fig. 12) is Pumping speeds for hydrogen and for
a mechanical vacuum pump that creates a helium vary only slightly from those for
gas flow toward a suitable forepump by air whereas the exhaust pressure is in the
imparting momentum or motion to gas range from 1.3 Pa to 1.3 mPa (10 mtorr to
molecules by means of a rapidly rotating 10 µtorr). Higher exhaust pressures are
rotor with successive rings with inclined achieved in compound turbomolecular
blades. These blades rotate with
circumferential speeds comparable to the FIGURE 12. Turbomolecular pump: (a) schematic; (b) inlet
thermal motion of the molecules (speeds port.
of 100 to 700 m·s–1 or 330 to 2300 ft·s–1).
Some molecules are struck by the rotor (a)
blades and rebound in a favorable axial
direction toward the stator blades. The Gas inlet
molecules rebound from these stator
blades in a direction favorable for their Power source
being impelled by the next stage rotor for motor
blades and so on as the process is repeated
through all successive stages of rotor and To forepump
stator blades. The series of impacts
statistically favor motion through the
turbine stages toward the discharge port
and constitute a pumping action with a
very high compression ratio.
The seal between the individual stages
is achieved by very narrow air gaps. The
FIGURE 11. Vacuum system with blower
pump.
Chamber Roughing valve
(b)
High Stator
vacuum
valve
Blower pump
Mechanical pump
Rotor
228 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pumps. These follow the turbomolecular These vapor streams are directed
stages with one or several molecular drag toward the outer walls of the pump. The
stages, which further compress the gas walls are typically cooled by water. When
through the effects of viscosity. the vapor hits the cooled walls, it
condenses back into a fluid. This fluid
Operation of Vapor or then flows downward into the pump
Diffusion Pumps for boiler for reboiling.
Vacuum Systems
The actual pumping of gases happens
Although mechanical rotary pumps when the large, heavy, high speed oil
sometimes attain pressure below 0.1 Pa vapor molecules hit gas molecules. The
(10–3 torr), they are generally used in the gas molecules are knocked downward and
100 to 0.1 Pa range. To obtain pressures compressed by the movement of the
well below 0.1 Pa, the vapor pump was at vapor jet stream.
one time the most commonly used.
However, in the 1990s it was largely The gas molecules are thereby
replaced by turbomolecular pumps compressed in several stages to higher
because of the backstreaming of vapors. pressures. They are finally pumped away
The principle of operation of vapor through the foreline by the mechanical
pumps is entirely different from that of a pump (Fig. 14). When the oil drops to the
rotary oil sealed pump, where the gases bottom of the pump, it is reboiled and the
and vapors are compressed by a rotating cycle repeats.
mechanical member and exhausted to the
atmosphere. Vacuum Limitations of
Vapor Diffusion Pumps
The vapor pump, or diffusion pump,
operates in the molecular flow region. The A diffusion pump (Fig. 14) cannot
basic principle involved is shown in operate at pressures above 0.1 Pa (1 mtorr)
Fig. 13. because the oil vapor jets cannot form in
the viscous flow region. Therefore, the
The pump works by heating the pump pump must start pumping in a chamber
fluid to its boiling point. The vapors that is already under vacuum (such as
travel upward inside the jet assembly and that attained with a rotary mechanical
exit through the jet nozzles. In fact, they forepump). Oil is the most frequently
are accelerated downward through the jet used diffusion pump fluid because of its
nozzles. The vapor molecules travel very low vapor pressure at room temperature.
fast and can reach supersonic speeds. Oil has a fairly steep curve relating its
pressure to temperature. This is necessary
FIGURE 13. Principle of operation of high vacuum vapor for proper operation of the pump boiler.
pump. Vapor forced through a narrow opening (nozzle) The lowest attainable pressure of the
attains a high speed and is directed at a downward angle. diffusion pump is determined in part by
Molecules of gas or vapor that wander along a path toward
the jet stream will be struck by vapor molecules. The gas FIGURE 14. Construction of three-stage high vacuum vapor
molecule B has diffused into the path of the jet stream where pump.
it is struck by the vapor molecule A. Molecule B is given a
generally downward motion. Cold cap Inlet Cylindrical
water cooled
Flow from Multistage body
vacuum chamber jet assembly
Exhaust
Nozzle Thermal Baffles
protect Foreline
switch
A Ejector
B Electrical connector
Fill and
drain
assembly
Vapor Vapor jet Oil reservoir (boiler) Heater
Leak Testing of Vacuum Systems 229
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
the vapor pressure of the oil at the of gas, it should not sacrifice high
temperature of the available cooling conductance because that would impair
water. Oils specified by pump the net pumping speed of the system.
manufacturers have vapor pressures,
under these conditions, of about 0.1 µPa Operation of Cold Traps in
(1 ntorr). The popularity of the diffusion Vacuum Pumping Systems
pump is due to its wide range of
operation, low cost, reliability and lack of A cold trap placed above the baffle
moving parts. The pump heaters are ensures that those few oil molecules that
usually mounted from the outside and may get by the baffle will not get to the
can be replaced during operation. vacuum chamber. A cold trap, therefore,
stops back migration of pump oil vapors.
Should the diffusion pump be suddenly It is also very effective as a cryogenic
exposed to a burst of atmospheric pump for pumping condensable vapors
pressure, the oil jet stream would collapse, such as water vapor, the chief offender in
thereby destroying the pumping most systems, as well as for grease vapors
capability of the vapor pumps and and other undesired contaminants.
possibly acting to crack the oil. The term
cracked oil refers to a decomposition of As a cryogenic pump, the cold trap
pump oil due to exposure to oxygen in reduces system pressure by taking
the atmosphere while at or near the molecules out of the gas or vapor phase
boiling point of the oil. Some fluids are and trapping them on its surface. These
less susceptible to cracking than are other molecules are not pumped out of the
diffusion pump fluids. vacuum system and discharged to
atmosphere. The most common
Operation of Baffles and techniques used to obtain low
Traps in Vacuum Pumping temperatures for cold traps are mechanical
Systems refrigeration, dry ice and liquid nitrogen.
One of the objections to diffusion pumps Some common forms of optically dense
has been the possibility of contaminating chevron and cold traps are shown in
the vacuum chamber work area with the Fig. 16, which also shows thimble type
pump fluid. By providing suitable traps traps used in mass spectrometer leak
and baffles between the pump and the detectors. The reservoir is filled with
vacuum chamber, back diffusion of oil liquid nitrogen through the filler tube.
and oil vapor can be minimized and Use of liquid nitrogen requires that the
condensable vapors from the chamber thimble type trap be kept essentially in a
may be trapped. As a general rule, the vertical position.
pumping speed of the system goes down
as the trapping efficiency of baffles and Characteristics Desired in Vacuum
traps goes up, due to decreased Valves
conductance.
Vacuum valves must (1) be free from
The baffle or trap should normally be leakage, (2) offer minimum flow
kept as cold as possible. However, the
temperature of surfaces of the first FIGURE 15. Typical baffle designs used in oil diffusion vacuum
baffling state above a pump should be pump systems.
cool enough to condense the oil vapors,
but not so cold as to freeze the pump oil Plate
and prevent it from flowing back into the
pump. Cooling coils
Characteristics Desired in Cooling
Diffusion Pump Baffles coils
A baffle is simply a cool surface that is Top view
placed above the diffusion pump in the of disk
path of gas flow. This baffle is of metal
with good thermal conductivity that Cooling
keeps its surface at a uniform temperature. coils
The refrigerant, usually cold water, is
passed through tubing that is brazed to Cooling coils
the baffle. A baffle should also be
optically dense, that is, there should be Cooling coils Top view
no line of sight through it, to avoid back of chevrons
flow of molecules in molecular flow.
Fig. 15 shows some typical designs of Cooled chevron trap
baffles. Because a baffle restricts the flow
230 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
resistance and (3) contain materials that cracking the bellows. Never completely
do not outgas. The biggest problem in extend bellows when out of the valve.
making leaktight valve is in sealing the
operating shaft. Two types of valves that Operation of Capture
accomplish efficient sealing are the Vacuum Pump
bellows sealed and diaphragm valves
(Fig. 17). Usually preferred are brass or Unlike the previously described pumps,
stainless steel bellows, more movement which compress and exhaust gas either to
being obtained with brass. The bellows is atmosphere or into an attached forepump,
brazed to the cover (bonnet) and dish, as two commonly used pumps collect and
shown in Fig. 17a. Figure 17b shows a store gases in the pump body until
valve using a diaphragm that can be a eventually being released to atmosphere
metal or elastomer. Compared to metal by a process called regeneration (for the
diaphragms, an elastomer has cryopump) or until the pump is rebuilt as
considerable flexibility but also has the in the ion pump. These pumps are the
disadvantages of outgassing and mechanical cryopump and the sputter ion
permeability to various gases. On the pump.
other hand, metal diaphragms are not as
elastic but have better outgassing and FIGURE 17. Operating principles of vacuum valves:
permeability characteristics. (a) bellows sealed valve; (b) diaphragm valve.
Precautions in Disassembly of (a) Cover
Bellows Sealed Valves
Always open bellows valve before
removing the stem assembly to prevent
FIGURE 16. Cold traps used in vacuum pumping systems to Bonnet gasket
condense vapor molecules: (a) combination baffle and trap
with optically dense chevrons; (b) thimble trap used in leak Braze
detectors. Bellows
(a) Body
Liquid Seat Braze
nitrogen Vent hole
Diaphragm
Water
(b) (b)
Liquid nitrogen
To Braze and mechanical seal
chamber if an elastomer diaphragm
To diffusion
pump
Leak Testing of Vacuum Systems 231
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Operation of Cryopump relief valve. In the expander, high pressure
helium is supplied by the compressor.
The cryopump is unique in that it pumps This gas is expanded in two stages to
by getting the gases so cold that they produce cryogenic temperatures. The
freeze and are stored, or captured, in the actual operating temperatures will vary,
pump. It is extremely clean, using no oil depending on the thermal and gas loads
and having no moving parts in vacuum. that are imposed. The first stage operates
It also has a very high throughput and is between 50 and 80 K (–370 and –315 °F)
used in the high vacuum range in and the second stage between 10 and
industrial applications where 20 K (–440 and –420 °F). The cryoarrays
hydrocarbons cannot be tolerated. are the pumping surfaces, cooled by the
expander, on which gases from the
A cryopump (Fig. 18)2 is made up of vacuum chamber are condensed or
two main components: a gaseous helium adsorbed.
compressor and a pump consisting of an
expander, cold head (chilled surfaces) and In cryopump operation, helium is
the pump body. These two components compressed and gives up its heat to the
are connected by flexible hoses to form a surrounding walls of the compressor. This
closed loop refrigeration system. Gaseous heat is removed by water or air cooling.
helium is circulated between the The cooled, compressed helium then goes
compressor and expander. to the pump cold head. The expander at
the cold head valving system lets the
The pump module consists of the helium expand. The expanded helium
expander module, the first and second now absorbs heat from the cold head and
stage cryoarrays, the pump body, second baffle array. This chills the cold head and
stage temperature monitors and a pressure baffle array to about 12 K (–440 °F) and
70 K (–335 °F), respectively. These chilled
FIGURE 18. Schematic of cryopump. surfaces pump gases from the vacuum
chamber in two ways. The gases are either
C condensed or adsorbed on the arrays.
That most gases will stick to a surface in
D an icelike state at less than 20 K (–420 °F)
E is very likely. At this temperature, the
combination of partial pressures of most
F gases is about 10–9 Pa (10–11 torr) or lower.
G Most gases are condensed on the first
and second stage cryoarrays. The first
H stage array is cold enough to pump water
I vapor and carbon dioxide by
cryocondensation. The colder second
A stage array pumps nitrogen, oxygen,
argon and most other gases by
J cryocondensation, but is not cold enough
to condense helium, hydrogen and neon.
B These three gases are pumped by the
process called cryosorption; a surface
Legend related phenomenon: the greater the
A = Forevacuum port available surface area at cryogenic
B = Power connection temperatures, the more likely that gas
C = Inlet flange molecules will stick to it. Although most
D = Baffle gases are frozen or condensed between
E = Second cold stage 12 and 20 K (21 and 36 °R), helium,
F = Radiation shield hydrogen and neon are still very actively
G = Cryoplates in motion at these temperatures. If we did
H = Relief valve not remove them, their partial pressures
I = First cold stage would continue to rise, perhaps to a point
J = Vapor pressure thermometer where the total system pressure would be
unacceptable.
To solve this problem, activated
charcoal is attached to the bottom side of
the second stage (coldest) cryoarray where
it is less likely to adsorb the easier to
pump condensible gases. This reserves the
charcoal for the helium, hydrogen and
neon which are trapped in the maze like
structures and surfaces of the charcoal.
This is similar to a sponge soaking up
water vapor at room temperature. This
process is called cryosorption.
232 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Before chilling the cryoarrays, the are accelerated toward the anode. This
pump volume must be rough pumped to long path increases the probability of
remove most of the gas below a valve at ionization and therefore the amount of
the pump inlet. During chilling, when the useful pumping action that can be
second stage drops to less than 20 °K, the performed by the pump.
pump is ready for use. During use, the
pump can absorb very large amounts of Because of the action of the magnetic
condensible gas, but the second stage field, the electrons do not easily come in
charcoal eventually saturates, usually with contact with the anode. As a result, a
hydrogen and must then be turned off to cloud of electrons is formed within the
warm up the pump (regeneration). To anode space. This electron cloud becomes
speed up this process, dry nitrogen is fairly stable during pump operation and is
applied to the purge tube, through a dense enough for the efficient ionization
valve, which flushes the pump and expels of gas molecules. The name for this
the previously stored gas out through the process is cold cathode discharge. The
pressure relief valve. When the second positively charged ions, which are
stage rises to room temperature, the pump relatively heavy particles, are accelerated
is ready to be rough pumped and chilled into the negatively charged titanium
again. cathodes. This impact causes sputtering,
or chipping away of the titanium cathode
The cryopump is normally used in the material.
pressure range of 10–1 to 10–6 Pa (10–3 to
10–8 torr) but when operated continually Sputtered titanium deposits onto the
at the upper end of this range, the time internal structure of the anode. Then,
between required regeneration cycles is when gas molecules come in contact with
proportionately shorter; i.e. — more these clean titanium deposits, chemical
downtime.
FIGURE 19. Section through a cold cathode ionization gage
Operation of Ion Pump (Penning gage).
The ion pump (Fig. 19)2 is also a gas AH
capture pump but is not designed to
pump heavy gas loads. For this reason, it B I
is not generally used alone in high J
production applications. It is more often C K
used in research and analytical
applications where there is no need to D L
cycle the work chamber repeatedly and E M
rapidly from atmosphere to vacuum. N
F
Ion pumps are clean operating
electronic devices which use no moving G
parts or oils within the vacuum pump. It
is possible to achieve pressures in 10–9 Pa Legend
(10–11 torr) range with overnight bakeout A = High voltage connection
of the system. The bakeout process drives B = Hood
residual gas off the system walls, which is C = Protective cap
then pumped by the ion pump. In D = Vacuum tight cast iron housing
research and analytical applications, the E = Permanent magnet
ion pump’s cleanliness, bakeability, low F = Small flange connection
power consumption, and long life make it G = Baffle
the pump of choice for most ultrahigh H = Safety terminal
vacuum uses. They are available in various I = Leadthrough (anode lead)
sizes and variations, but only the simplest J = Compressed glass-to-metal seal
(diode) pump will be described here for K = Ring anode
purposes of brevity. L = Ignition pin
M = Fixing screw
A stainless steel ion pump body N = Cathode plate (exchangeable)
contains a multicell anode assembly
constructed of cylindrical parallel tubes
spaced between two flat titanium
cathodes. A very strong magnet is placed
outside the pump body. After the ion
pump is rough pumped to 1 Pa (10–2 torr)
or less, a voltage of 5 to 7 kV direct
current is applied between the cathodes
and the anode assembly.
The magnetic field forces any free
electrons within the anode into long
helical paths instead of straight paths.
This increases the probability of electron
collision with molecules, as the electrons
Leak Testing of Vacuum Systems 233
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
combination converts these gas molecules the system. At lower pressures, ion pumps
to solid compounds such as titanium have long lives. Once they begin
oxide or titanium nitride. This process is pumping, they quickly lower the pressure
called chemical gettering and produces to the long life region. As long as they are
the required pumping action. In addition, not pumping against a leak, they will last
a second pumping action takes place. for years. An example of this would be
Some of the ionized molecules impact the that a pump working at a constant
cathodes with enough force to become pressure of 10–5 Pa (10–7 torr) would have
buried in them, which prevents them a useful life of 20 years.
from neutralizing and becoming a free gas
again. A third pumping action occurs Procedures for Pumping
with hydrogen which diffuses directly and Operating Complete
into and reacts with the cathode plate. Vacuum Systems
Also, neutral particles such as the inert
gases can literally be buried or covered by By combining the components previously
the sputtered cathode material. Complex discussed with appropriate manifolding,
molecules may also be split apart in the plumbing and gaskets (O-rings), a
discharge to smaller, more readily pumped complete vacuum system may be built as
molecules. Because these actions are not shown schematically in Fig. 20. The initial
equally efficient, the chemically reactive conditions are:
gases such as hydrogen, nitrogen and
oxygen are pumped at much higher 1. mechanical pump running,
speeds than the inert gases. A 2. diffusion pump operating and working
modification of the cathode design can be
made to increase the efficiency for these in high vacuum,
inert gases. 3. cold trap filled with liquid nitrogen,
4. atmospheric pressure in bell jar
Another characteristic of the ion pump,
often referred to as a sputter ion pump, is chamber,
that it is self-regulating. At higher 5. high vacuum valve closed,
pressures, where much ionization takes 6. vent valve open,
place, more current flows and at low 7. roughing valve closed and
pressures, less current flows. This 8. foreline valve open.
characteristic current drain can be used to
measure the pressure, or degree of vacuum An operational cycle for this vacuum
achieved with the pump. This feature system is as follows:
eliminates the need for an ion gage on
1. Close access to bell jar chamber, vessel
FIGURE 20. Schematic diagram of a typical complete vacuum or hood to be evacuated.
system.
2. Close vent and foreline valves. The
Pumping Bell jar chamber ballast tank permits the
port or turbomolecular pump or diffusion
other process vessel pump to discharge to an expansion
Vent valve volume so that a high critical
forepressure is not reached.
High Roughing
vacuum valve 2 3. Start the roughing cycle by opening
valve 1 the roughing valve. This allows the
Cold trap Foreline mechanical pump to evacuate the
Diffusion manifolding between the high vacuum
pump Baffle valve 3 valve and the bell jar chamber.
To 4. After the pressure has been reduced to
atmosphere below 10 mPa (about 50 µtorr) or
crossover point, close the roughing
Ballast valve.
tank
5. Open the foreline and high vacuum
Leak test valves. This allows the diffusion
port pumping system (cold trap, baffle and
turbomolecular diffusion pump) to
Mechanical pump continue pumping until the desired
operating pressure is reached and work
in the chamber may commence.
After completion of work in the bell jar
or vacuum chamber, the system may be
cycled to its initial condition by first
closing the high vacuum valve and then
opening the vent valve. This allows
atmosphere to enter the bell jar chamber
and system up to the roughing and high
vacuum valves. The pressure equalization
allows access to the chamber.
234 Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 3. Materials for Vacuum Systems
Outgassing of Materials in This set one limit on the lowest ultimate
Vacuum Systems pressure that can be reached in a
particular vacuum system. The usual
Adsorption refers to the condensation of technique of overcoming this problem is
gas (vapor) on the surface of a solid. As to degas the materials, usually by baking
the pressure is reduced in a vacuum (raising the system to a high temperature
chamber, there is a spontaneous evolution while pumping). The bake-out
of gas (and vapor) from materials in the temperature will depend on the
vacuum; this is referred to as outgassing. temperature at which the material begins
to change its properties. consequently,
In vacuum systems the materials in the vacuum systems are degassed at fairly
vacuum region may release adsorbed gases modest temperatures, say 300 to 400 °C
and vapors that increase the gas load of (570 and 750 °F), for several hours while
the system, resulting in a much longer being pumped. This will eliminate much
pumpdown time. This phenomenon is of the adsorbed gases and vapors.
most prevalent in new vacuum systems,
unclean vacuum systems or vacuum The dissolved gas content of a metal or
systems that have been exposed to alloy will depend on factors such as
atmosphere for some considerable time. It (1) the nature of the metal, (2) the
will also occur when new materials or metallurgical process used in the
new work jigs and fixtures are installed in production of the metal and (3) the
a vacuum chamber. Knowledge of the gas degreasing and cleaning to which a metal
adsorption properties of various materials was subjected. In comparing metals such
and, therefore, their outgassing properties, as stainless steel and aluminum, stainless
is very valuable in vacuum work. steel is found to outgas at a much lower
rate. A cast aluminum surface outgasses at
Technique for Releasing a rate about ten times higher than the
Adsorbed Gases by rate at which a stainless steel surface
Moderate Heating outgasses. Therefore, during vacuum
pumping, stainless steel vacuum systems
Most metals in vacuum give off are capable of reaching a desired vacuum
adsorbed or dissolved gases as well as in a shorter time than a comparable
gases resulting from the decomposition of aluminum system with its higher rate of
oxide near the surface. To minimize this outgassing. Results are strongly influenced
gas evolution, metals can be heated under by the condition of a metal (its alloy,
vacuum before being used in vacuum cleanliness, finish etc.).
systems. Gas adsorbed by exposure to
atmospheric pressure can easily be released Functions of Elastomers as
by heating to moderate temperatures. Gaskets and Seals in
When pumping to pressures below 0.1 mPa Vacuum Work
(1 µtorr) where baking is not practical, great
care must be taken in choosing the various Certain openings must be provided for the
materials in the system. This applies to insertion, removal and sealing of
choice of vacuum greases, elastomers, equipment or materials for a given
metals and various sealing compounds. vacuum system. During the operation,
these openings must be tightly sealed.
Factors Influencing Elastomers are the most widely used gasket
Adsorption and material, where temperature and gas loads
Outgassing by Baking permit, because they offer reliable sealing.
Vacuum System Materials Elastomers are natural or synthetic rubbers
that can be vulcanized to a state in which
Gases and vapors are adsorbed by vacuum they have an inherent ability to accept
construction materials (metals and and recover from extreme deformation.
elastomers) and are gradually released.
For high vacuum service, leaks must be
entirely eliminated and the gas evolved
from the gasket material itself must be
negligible. Both natural and synthetic
rubber satisfy these requirements as long
Leak Testing of Vacuum Systems 235
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
as negligible surface area is exposed. room temperature and at pressures near
Frequently, gaskets must be exposed to oil 1 mPa (10 to 1 µtorr). Because of its
or other gasket deteriorating substances temperature tolerance, silicone rubber is
and sometimes rather high or low commonly used for low and high
temperatures must be tolerated. Gasketed temperature operation. The fluorinated
joints should be readily accessible for tests elastomers are highly resistant to most
for leakage. In designing a gasket these corrosive materials found in vacuum
factors must be considered and practice. Fluorocarbon resin is very good
specifications should be based on material but suffers from cold flow under pressure
capabilities as well as vacuum system at room temperature; suitable means for
operational requirements. containing the fluorocarbon resin (spring
loaded gaskets etc.) will eliminate this
Selection of Gasket difficulty. In trying to reach very low
Materials and Design for pressures, the permeability of the elastomer
Vacuum Seals as well as its outgassing characteristics must
be considered. Permeability is the property
The choice of natural or synthetic rubber that determines how readily gases will pass
for a vacuum application depends on the through a material.
combined qualities desired. In the case of
rubbers, a wide range of characteristics is Selecting Elastomers to Reach Low
acceptable. Perhaps the most important Pressure Vacuums
single factor is that of allowable deflection
under compression. This is a function of To reach low pressures at room
hardness and allowable permanent set. temperature, elastomers with low vapor
These materials generally contain volatile pressures and low permeabilities are
oils, plasticizers and coloring pigments desirable. consequently, considerable work
that adsorb moisture and gases. Most of has been done with fluorinated elastomers.
the chemicals used have low vapor Baking an elastomer at a temperature that
pressure at room temperature. The does not damage it will reduce pumpdown
outgassing rates for various elastomers time; however, it will still release some
depend on factors such as (1) the vapor after many hours of pumping.
formulation used, (2) the area exposed,
(3) the operating temperature and (4) the Selection of Alloys for Use
treatment of the elastomer before use. As in Vacuum System
a rule, there is no way to control the Components
formulation of gasket materials because
this is determined by the manufacturer. There are many alloys of copper, but only
However, it is feasible to inform the brasses and bronzes are used in vacuum
manufacturer of intended service and ask practice. Brasses are copper zinc alloys,
for minimum volatiles. Exposed gasket whereas bronzes are copper tin alloys.
area becomes critical as the operating However, many brasses contain various
pressure is lowered. Proper gasket groove other metals. Brasses are widely used for
design can help considerably in reducing vacuum parts, such as diffusion pump
exposed areas. parts, chambers, base plates, valves and
fittings in high speed dynamic vacuum
Because the outgassing rate of systems.
elastomers increases as the temperature is
raised, the ultimate pressure can be Many commercial bronzes contain
reached more rapidly if the elastomer can zinc. Alloys containing zinc, cadmium,
be heated. However, all elastomers are lead, antimony or bismuth should not be
damaged when heated too much. Also, used in vacuum systems that are to be
the compression set increases more baked because of the high vapor pressures
rapidly with temperature. of these metals. Vacuum firing is likely to
alter the composition and therefore the
Because of these properties, elastomeric properties of such alloys.
gaskets are not normally used in ultrahigh
vacuum systems. Such systems are baked Properties of Austenitic Stainless
at temperatures well above the damage Steels in Vacuum Systems
point of all known elastomers. In this
case, it becomes necessary to use joints Stainless steels have come into fairly
and seals of metals and alloys such as common use in vacuum practice for
aluminum, brass, bronze, copper, indium, turbomolecular pumps, diffusion pumps,
lead, silver, stainless steel and others. manifolds, chamber baseplates etc.
Austenitic stainless steels (types 302, 303
Properties of Specific Elastomers and 304) are commonly used in vacuum
for Vacuum Seals work and are often called 18-8 stainless
steels because they contain about
Natural and synthetic rubbers are 18 percent chromium and 8 percent
commonly used in systems that operate at
236 Leak Testing
pyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
nickel. These steels are nonmagnetic and Selection and Properties of
the melting points of austenitic stainless Vacuum Greases and Oils
steels are over 1400 °C (2550 °F). Surfaces
of stainless steels remain smooth because Vacuum greases are commonly used to
oxides and hydroxides do not occur as in help attain seals and to lubricate devices
other types of metals. This means that the such as stopcocks and gasketed joints
effective surface area is less and vapors are (static, rotating and sliding). In some
adsorbed in smaller quantities. This leads cases, vacuum oils are used, including
to much easier degassing and quicker diffusion pump oils. Oils are generally not
pumpdown. as satisfactory as greases for most types of
seals, because they are more readily
Properties of Aluminum Alloys squeezed out, thereby leaving a dry seal.
Used for Vacuum System In general, vacuum greases should not
Components have a vapor pressure of more than about
10 mPa (0.1 mtorr) at 30 °C (86 °F) and
Aluminum is also being used in vacuum should maintain adequate viscosity at this
systems. The alloys of aluminum are temperature and can be used up to a few
generally readily worked in the shop degrees below their melting point. In
without much difficulty, the workability general, vacuum greases should be applied
depending on the composition. Surface sparingly and surplus grease then wiped
hardening can be achieved easily by off, because greases absorb gases and
anodizing and other processes. Parts may vapors and are dirt catchers.
be joined together by using aluminum
solder. Cast aluminum alloy parts are used Diffusion Pump Oils
for a variety of purposes such as valves,
turbomolecular pumps, diffusion pumps The ultimate vacuum of many vacuum
(particularly jet assemblies), grooveless systems is, in fact, limited by insufficient
flanges and gaskets. The design of the dies trapping of gas molecules by the diffusion
is important to get vacuum tight pump fluid. Certain desirable properties
aluminum die castings. Although that a diffusion pump oil must have
aluminum is difficult to de-gas include the following.
thoroughly, it is commonly used for
vacuum parts where good heat and 1. It should have low vapor pressure.
electrical conductivity is required. Vapor pressures of typical diffusion
pump oil recommended by
Properties of Other Metal Seals in manufacturers of diffusion pumps are
Vacuum Systems in the range from 10 to 0.01 µPa (100
to 0.1 ntorr).
Certain specialty metals have almost the
same coefficient of expansion as most 2. It should have low enough viscosity to
glasses and have excellent sealing flow back into the boiler.
characteristics. They are used with
vacuum flanges in the manufacture of 3. It should have high molecular weight
ionization gage tubes and in other relative to the pumped gases to
applications where metal-to-glass increase the efficiency of removal of
junctures and seals are necessary. gas from systems being evacuated by
the vapor jets. Molecular weight of oil
Applications and commonly used is in the range of 300
Limitations of Soft Metallic to 500 unified atomic mass units (u).
Vacuum Gaskets
4. Oil should be thermally stable to
Metal gaskets of some kind are used by avoid decomposition with heat.
vacuum seals that must be maintained at Decomposition often results in the
temperatures higher than about 125 °C evolution of more volatile fractions
(257 °F) or in which rubber cannot be caused by cracking of the oil due to
used because of outgassing. Small gaskets frequent exposure to atmospheric
of lead, copper, aluminum, gold, silver or pressures.
tin have long been used for higher
temperature vacuum services. Complete 5. The fluid should be chemically stable
sealing demands high stresses and and noncorrosive in the presence of
consequently the metal gaskets can only common metals, glass, elastomer
be used once. They are not designed for gaskets and the gases and vapors
applications where the seals are often usually present in vacuum systems.
opened and then reclosed because the
metal gaskets will take a permanent set 6. It should be nontoxic.
and are not reusable in most applications.
The recommended hydrocarbon oils
represent a very satisfactory low cost fluid
for the normal vacuum range down to the
low 10 µPa (100 ntorr) region. Ultrahigh
vacuum is best obtained with oils specified
by pump manufacturers. These oils are
extremely stable, showing little change in
properties even if the pump is exposed to
atmospheric pressure with the heater on.
Leak Testing of Vacuum Systems 237
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The words you are searching are inside this book. To get more targeted content, please make full-text search by clicking here.
ASNT NDT Handbook Volume 1_ Leak Testing
Discover the best professional documents and content resources in AnyFlip Document Base.
Search
ASNT NDT Handbook Volume 1_ Leak Testing
- 1 - 50
- 51 - 100
- 101 - 150
- 151 - 200
- 201 - 250
- 251 - 300
- 301 - 350
- 351 - 400
- 401 - 450
- 451 - 500
- 501 - 550
- 551 - 600
- 601 - 650
- 651 - 652
Pages: