ASM Metals HandBook Vol. 17 - Nondestructive Evaluation and Quality Control
Подождите немного. Документ загружается.
Because mass spectrometers must operate in a vacuum, they are ideally suited to the leak testing of vacuum systems.
They can also be adapted to testing systems pressurized with a tracer gas by using either a detector probe or enclosure to
collect the leakage while monitoring with the mass spectrometer. When used to test a pressure system, the mass
spectrometer is much less sensitive and the minimum detectable leak rate is approximately 10
-4
atm cm
3
/s. Special
assemblies containing a pump and permeation membrane, which are placed between the detector probe and the instrument
manifold inlet, can decrease response time and increase the sensitivity for detector probe tests of pressure systems to 10
-6
atm cm
3
/s. Small units, such as sealed electronic components or nuclear-fuel elements, are often internally pressurized
with tracer gas; leakage is detected by placing these units in an evacuated bell jar and monitoring the outgassing with a
mass spectrometer (Fig. 7). In general, the basic methods used in mass spectrometer leak testing include detector probe
(sniffer) testing, encapsulator testing, tracer probe testing, hood testing, bell jar testing, accumulation testing, pressure
vacuum testing, and differential sorption testing.
Fig. 7
Schematic of the leak testing of sealed components internally pressurized with helium tracer gas and
enclosed in a bell jar
Radioisotope Testing. A number of leak testing techniques have been developed that involved the use of solutions
containing radioactive tracers. In these tests, water is the liquid most commonly used, although hydrocarbons are
extensively used in the petroleum industry. Soluble salts and compounds containing radioisotopes are added to the liquid.
One of the most satisfactory isotopes that can be added to water is the sodium isotope
24
Na. The use of this radioisotope is
safe because the concentration required for the detection of leaks is less than the drinking tolerance. Furthermore, this
radioisotope has a short half-life (15 h) and can be retained in a vessel until its activity has been greatly reduced as a result
of radioactive disintegration (decay).
Using this method, the vessel is filled with radioactive sodium bicarbonate solution and pressurized to the level required
for a code proof test. After the vessel is held at pressure for a short period of time, it is drained and flushed with fresh
water. A gamma-radiation detector is used to determine the location of leaks from the presence of gamma radiation
emitted from the radioactive sodium bicarbonate solution that has accumulated around leaks in the vessel.
In another method, small components are backfilled with
85
Kr by placing them in a chamber and then pressurizing the
chamber with
85
Kr (Fig. 8). If a component has a leak, gas will enter it. After the backfilling is complete, the components
are air washed, and the leakage is measured with a radiation counter. Under carefully controlled conditions, the radiation
counter has a sensitivity of about 10
-12
atm cm
3
/s. The radiation counter is automated and enables rapid, sensitive testing
of large quantities of small parts such as microcircuits.
Fig. 8 Schematic of basic steps in the
85
Kr tracer-gas bombing of hermetically sealed electro
nic components.
(a) Load parts. (b) Evacuate and remove air. (c) Soak parts in radioactive gas.
(d) Recover radioactive gas. (e)
Return air
One disadvantage of this method is that organic coatings on the parts may absorb the
85
Kr, thus giving a false leak
indication. The primary disadvantage of this method is the attendant health hazard. However, some pressurized systems
such as pipelines, lend themselves to leak testing by this method.
Ionization Gage Testing. The Phillips vacuum gage is used to measure the pressure of a gas by the intensity of a gas
discharge in a magnetic field. However, this form of gas discharge ceases when the gas pressure is reduced to a pressure
of approximately 0.13 Pa (10
-3
torr); therefore, the Phillips gage cannot be used for making measurements at low
pressures.
Penning Gage. A vacuum gage based on the same principle as the Phillips gage, but useful at lower pressures, was
constructed by replacing the electrode system with a system consisting of two parallel round metal disks (acting as
cathode and anode) enclosed in a cylindrical glass jacket. To keep the distance between the pole shoes of the permanent
magnet enveloping the glass tube as short as possible, the tube was punched at that point. This improved instrument is
called the Penning gage, and it has a sensitivity about ten times greater than the original Phillips gage. The Penning gage
covers a pressure range from 13 to 0.13 mPa (10
-4
to 10
-6
torr) and can probably be used even for pressures below 13 Pa
(10
-7
torr).
In any pumping installation, the vacuum gage can also be used to detect and locate leaks. This is done by replacing
atmospheric air with tracer gas in each part of the apparatus. When the leak is reached, the tracer gas flows in, and the
gage usually registers any change in pressure.
Additional Ionization Detectors. In addition to the Phillips and Penning gages, other ionization detectors include
the oxygen-leak detector, the hot-filament ionization gage, the helium-ionization leak detector, the photometric leak
detector, and the palladium-barrier detector.
Gas chromatographs are sometimes used for leak detection. In a gas chromatograph, a carrier gas is made to flow
through an adsorbent column; then a tracer gas is injected into the carrier gas ahead of the adsorbent column. Next, a gas
sample is fractionated into its constituents as a function of the residence time of each constituent in the adsorbent column.
The constituents are detected as they leave the adsorbent column, usually by thermal-conductivity instruments, but they
may also be detected by infrared analyzers or vapor-density balances. The advantage of the gas chromatograph is that,
unlike the mass spectrometer, with proper calibration it will identify the tracer gas as a compound rather than as
individual molecules. This can be beneficial if there are several gases in the system; their identification will often aid in
locating the source of the leakage. Residual gas analyzers and partial pressure analyzers are used in the same manner.
Specific-Gas Detector Applications
The proper method for using a specific-gas detector is based on the function of the leak detector, the fluid that is leaking,
and the type of vessel being tested. Some of the more important considerations in determining the best method of using
leak detectors are discussed below.
Probing. Detector probes, which will react to a number of gases, can be used in either of two ways. The first is the
scanning mode, and the second is the monitoring of an enclosure placed around the pressurized item (Fig. 9).
Fig. 9 Tracer-gas detection techniques used to locate leaks with sensitive electronic instruments.
(a) Tracer
probe. (b) Detector probe
Detector Probe Technique. In the detector probe mode, the external surface of the pressurized vessel is scanned
either with a portable detector having a short probe attached or with a long probe connected to a stationary leak detector
by flexible tubing (Fig. 10). In general, the connection of a long probe to a stationary detector reduces sensitivity because
of the slow release of sorbed gases in the probe tubing, which results in a high background reading. Because a
correspondingly longer time is required for the gas to flow up the tube to the sensing element, it is difficult to pinpoint the
location of the leak. Also undesirable is the long time needed for purging a long-probe system of tracer gas. The purging
occurs when a rather large leak is encountered and the probe tubing becomes filled with a high concentration of tracer
gas. One minute or more may be required for the leak detector pump to clear the tracer gas from the sensing element. This
in turn reduces scanning speed.
Fig. 10
Schematic showing a long detector probe connected by a flexible hose to a stationary leak detector for
scanning the external surface of a pressurized vessel
There are means for offsetting some of the disadvantages of the long-probe method. When a helium mass spectrometer is
used, carbon dioxide can be injected into the probe line near the tip. The carbon dioxide will act as a carrier fluid for the
helium, thus purging the tube walls, and, in conjunction with a high-capacity pump, will carry the helium tracer gas to the
leak detector very rapidly. The carbon dioxide is then selectively trapped with liquid nitrogen, leaving the helium. A
probe of this type is several times faster than conventional probes and reduces background buildup.
When background buildup and sensitivity losses do occur, accurate measurements of leak rate are difficult to obtain.
However, leak-rate measurement can be accomplished by placing a calibrated leak over the probe tip to calibrate the leak-
detector-tube-probe system. The leakage can then be measured by placing a suction cup on the probe and applying it to
the leak area.
Another, more recent, means of decreasing the disadvantages of the long-probe method is the use of an assembly between
the probe and the mass spectrometer manifold port, which contains a permeation membrane and pump. The pump reduces
the response time, and the permeation membrane reduces the pressure drop into the mass spectrometer and also causes
accumulation of the entering tracer, which increases sensitivity.
Another disadvantage of the long-probe method is that the operator generally cannot observe the reaction of the leak
detector. This can be circumvented with earphones, audible signals, and lights, but such devices must be triggered at some
threshold level of detector response that may not correspond to the desired leak size. One person can operate the probe
while another monitors the leak detector readout, but ideally one person should perform both tasks simultaneously.
Detector Probe Accumulation Technique. The second method of detector probing involves the use of a tracer-gas
leak detector to monitor an enclosure that accumulates the leakage (Fig. 11). This method does not provide a means of
locating the leak, but it is preferred for leakage measurement and, in many applications, for detecting the existence of a
leak.
Fig. 11
Schematic of setup for detecting leaks by monitoring an enclosure placed around a vessel pressurized
with tracer gas
A large unit can be tested by enclosing it in a plastic bag. The accumulation of tracer gas in the bag is monitored by a leak
detector connected to the bag by a short probe or by piping. Leakage measurement consists of inserting a calibrated leak
into the enclosure and comparing the test-leak rate with the calibrated-leak rate. One of the problems associated with this
method is the construction of a bag that does not leak and has low permeability with respect to tracer gas.
A small unit can be tested by placing it in an enclosure such as a sealed bell jar, which is monitored for the accumulation
of tracer gas. Depending on the type of leak testing instrument used, the enclosure may be placed under a vacuum. If
relatively large leaks are expected, a supply of tracer gas must be connected to the test unit to ensure that the
concentration of its internal tracer gas and the pressure drop across the leak are constant.
Back pressuring is a method of pressure testing that is normally used with small, hermetically sealed electronic
components, such as integrated circuits, relays, and transistors. In this method, the test unit is placed in a pressurized
container filled with a tracer gas and is kept there for a time to allow tracer gas to flow into the unit through any leaks that
may be present.
In using back pressuring, care must be taken to ensure that the leaking components contain tracer gas. Therefore, it is
important that the time between back pressuring and leak testing be suitably controlled. For example, in a test
specification for transistors, the transistors were subjected to a helium pressure of 690 kPa (100 psi) of 16 h, air washed
for 4 min, and then leak tested within 3 h. If more than 3 h had elapsed before leak testing, the transistors were
pressurized again. If there is doubt as to how much tracer gas will flow into the components under back pressurization,
the following expression can be used for estimation purposes:
(Eq 7)
where P
i
is the partial pressure of tracer gas inside the component (in atmospheres), P
o
is the pressure of the tracer gas
surrounding the component (in atmospheres), A is the conductance of the leak (in cm
3
/s) with a pressure differential of 1
atm, V is the internal free volume of the component (in cubic centimeters), and T is the time that pressure P
o
is applied (in
seconds).
Because leaks are generally expected to be quite small, very sensitive tracer-detector combinations must be used. The
tracer gas normally used is helium or
85
Kr, and detection is accomplished with a helium mass spectrometer or a nuclear-
radiation detector, respectively. Absolute values of the size of the leak are sometimes difficult to determine with back
pressuring because of:
• Adsorption and absorption of the tracer gases
• Different detector response times for different leak directions.
When testing complex systems, each sealed component can be pressurized with a different tracer gas. Leak detection with
a residual-gas analyzer will both detect leakage and provide a determination of which components are leaking.
Leak Testing
Revised by Gerald L. Anderson, American Gas and Chemical Company
Leak Testing of Pressure Systems by Quantity Loss
Determining the amount of fluid lost from a vessel is one means of establishing that leakage has occurred. Weighing a
vessel to determine fluid loss is practical only if the vessel is of reasonable size and can be weighed on a suitable scale.
The sensitivity and accuracy of this method are limited by those of the scale; in general, this method is unsatisfactory
unless the weight of the vessel is comparable to or less than the weight of its contents.
Differential-Pressure Method. Gas-quantity loss can be determined by the differential-pressure method. This is done
by connecting two identical containers together with a differential-pressure gage between them, as shown in Fig. 12.
Differential-pressure gages are sensitive to small pressure differences; therefore, if the two containers are maintained at
equal ambient conditions, leaks as small as 10
-4
atm cm
3
/s can be detected in systems operating at pressures up to 40 MPa
(6 ksi). The primary advantages of this method are that temperature effects are largely cancelled and the actual working
fluid can be used.
Fig. 12 Schematic showing the use of a differential-
pressure gage between a reference vessel and the test
vessel to detect leaks in the test vessel
Leak Testing
Revised by Gerald L. Anderson, American Gas and Chemical Company
Visual Leak Testing of Pressure Systems
The simplest method of checking a liquid-filled vessel for leaks is visual observation. Like listening and smelling, this is a
quick and simple method of testing for gross leaks in liquid-filled containers.
Vision is a variable commodity when considered from the viewpoint of a single individual and is even more variable
when considered from the viewpoints of several individuals. This is because variations in the eye, as well as variations in
the brain and nervous system, affect both optical resolution and color perception.
Optical aids to vision, such as mirrors, lenses, microscopes, borescopes, fiberoptics, and magnifiers, compensate for
some of the limits of the human eye by enlarging small discontinuities. For example, borescopes permit direct visual
inspection of the interior of hollow tubes, chambers, and other internal surfaces.
Hydrostatic leak testing requires that a component be completely filled with a liquid, such as water. The normal
sensitivity for visual inspection using deionized water as the test fluid is 10
-2
atm cm
3
/s. Therefore, only large
discontinuities can be revealed with this method.
The sensitivity of the test can be improved by the addition of:
• A water developer applied to the outside that changes
color when contacted by a small leak. This
method has replaced the older lime wash method, which required a larger amount of water seepage to be
visible
• A concentrate that lowers the surface tension of the water and provides a visible or fluorescent tracer
When conducted properly, the hydrostatic test can achieve the same sensitivity as a liquid penetrant leak test on large-
volume specimens. The penetrant test is two orders of magnitude more sensitive on volumes of typical electronic
components.
A recent variation of the hydrostatic test is the stand-pipe test, which is used to detect leaks in underground storage tanks.
In this test, the storage tank is overfilled, and the liquid level is monitored by reference to an aboveground liquid column
or stand pipe. This method is not very effective and has a sensitivity of between 10
-1
and 10
-2
atm cm
3
/s.
Detection by Water-Soluble Paper With Aluminum Foil. One method of detecting water leakage employs a strip
of aluminum foil laid over a wider strip of water-soluble paper. The strips are then laid over the welded seams of a water-
filled vessel, as shown in Fig. 13. If a leak exists, the water-soluble strip will dissolve, indicating the leak location, and the
aluminum foil strip will be in electrical contact with the vessel. A corresponding change in resistance indicates that a leak
is present.
Fig. 13 Cross section of a welded seam in a water-filled vessel showing strips of water-
soluble paper and
aluminum foil used to detect leaks in the seam
Liquid Penetrants. The use of a visible liquid penetrant for leak testing is only slightly limited by the configuration of
the item to be inspected. Another advantage is that this method can be used on subassemblies prior to completion of the
finished vessel. Leak testing with penetrants is equally suited to ferrous, nonferrous, and nonmetallic materials, provided
the materials are not adversely affected by the penetrant. Leak testing by this method cannot be substituted for a pressure
test when the applied stress and associated proof-test factors are significant.
The liquid penetrant should be applied in a manner that ensures complete coverage of one surface without allowing the
penetrant to reach the opposite surface by passing around edges or ends or through holes or passages that are normal
features of the design. A sufficient amount of penetrant should be applied to the surface during the period of the test to
prevent drying. Immediately after application of the penetrant, the developer should be applied to the opposite surface.
The developer must be prevented from contacting the penetrant at edges or ends or through designed holes or passages so
that the test results are valid.
If greater sensitivity is desired than is provided by a color-contrast (visible) liquid penetrant test, fluorescent liquid
penetrant can be used to increase the sensitivity approximately one-hundredfold. Additional information on the use of
liquid penetrant for leak testing is available in the article "Liquid Penetrant Inspection" in this Volume.
Smoke bombs and candles can also be used for detecting leaks. Generally speaking, a volume of smoke sufficient to
fill a volume five or six times larger than the area to be tested is required for a smoke test. Medium-size volumes, such as
boilers and pressure vessels, can be tested by closing all vents, igniting a smoke candle or smoke bomb, and placing it
inside the vessel. All openings should then be closed; almost immediately, escaping smoke will pinpoint any leaks within
the sensitivity of this method. When testing boilers and similar equipment with volumes of 2800 m
3
(100,000 ft
3
) and
more, it is often desirable to apply air pressure.
Smoke candles can provide from 110 m
3
(4000 ft
3
) of smoke in 30 s to 3700 m
3
(130,000 ft
3
) in 2 to 3 min. The smoke
will vary in color from white to gray, depending on density and lighting. It is generated by a chemical reaction, contains
no explosive materials, and is nontoxic.
Leak Testing
Revised by Gerald L. Anderson, American Gas and Chemical Company
Leak Testing of Vacuum Systems
Vacuum systems are normally tested in the evacuated condition, rather than being internally pressurized. Because some
leaks, particularly small ones, are directional, it is desirable to test with the leakage flowing in the operational direction.
Introduction of Tracer Gas Into System
There are two ways in which a tracer gas can be used. One is to tracer probe to both detect and locate the leak by spraying
tracer gas over the outer surface of the vessel. The other is to envelop the entire test vessel (or a particular portion of it) in
a bath of tracer gas by the use of either rigid or flexible enclosures to both detect and measure leak rate. Calibration is
accomplished by injecting a known quantity of tracer gas into the vacuum system or by attaching a calibrated leak to that
system.
Helium is one of the most commonly used tracer gases for the following reasons:
• It is basically inert and safe
• It has a small molecular size and thus can flow through very small leaks
• Its diffusion rate is high
• It is easily detected by a mass spectrometer
In testing a vacuum system, even the smallest leaks are important because they have a great influence on the ultimate
vacuum that the system can reach. Therefore, the most sensitive test method--helium tracer gas in combination with a
helium mass spectrometer--is normally used.
Certain characteristics of high vacuums and high-vacuum systems must be recognized and understood prior to testing
such a system for leaks. These characteristics are briefly discussed below.
Tracer-gas diffusion rates are highly variable and are dependent on the type of gas used and the configuration of the
system. The diffusion rates of such gases as helium and hydrogen are generally eight or nine times higher than those of
the Freons.
Attachment of Mass Spectrometer. Mass spectrometers incorporate pumping systems that can be used to maintain
the vacuum if small systems are being tested and the leakage is small. With large systems, it may be necessary to operate
system diffusion vacuum pumps backed with forepump mechanical vacuum pumps while the system is being tested. In
general, the mass spectrometer should be attached to the foreline of the diffusion pumps, parallel with the forepumps, and
not to the test vessel (Fig. 14a). In this manner, the mass spectrometer will receive an amount of tracer gas that is
proportional to the pumping rate of the spectrometer and the forepump according to:
(Eq 8)
where T
G
is the percentage of total tracer-gas leakage flowing to the mass spectrometer, S
1
is the spectrometer pumping
speed, and S
2
is the forepump pumping speed. If the mass spectrometer is connected directly to the test vessel, it will
receive only a small amount of tracer gas because the spectrometer pumping speed is generally far lower than that of the
vacuum chamber diffusion pumps (Fig. 14b). Conversely, when the mass spectrometer is connected to the diffusion pump
foreline, the compression of the evacuated gas increases the total gas pressure in the foreline and in turn increases the
tracer gas partial pressure in that line, which makes it more detectable than if it were sensed directly from the vessel. If
the system throughout is small, however, the system diffusion pumps and roughing pumps can be valved off or backed by
the mass-spectrometer pumps, which in turn will maintain the desired vacuum.
Fig. 14 Schematics
of correct (a) and incorrect (b) connections in pumping setups for the vacuum leak testing
of large volumes
System Materials and Cleanliness. Both the vacuum system and the leak testing system should be maintained in a
very clean state; otherwise, contaminant outgassing will result, reducing the system vacuum. The proper materials must
also be selected to reduce hang-up of the tracer gas; glass or stainless steel is recommended. Grease coatings on seals and
the use of rubber should be held to a minimum because these materials readily absorb helium and then slowly release it,
giving a high background count.
Temperature and Tracer-Gas Concentration. For leak testing, the temperature should be held constant, when
possible, preferably at the expected operating temperature of the item tested. Also, tracer-gas concentration should be as
high as possible to ensure sufficient detector response.
Vacuum Leak Testing Methods
The equipment most commonly used for the leak testing of vacuum systems involves the use of manometers, halogen
detectors, mass spectrometers, ionization gages, thermal-conductivity gages, residual gas analyzers, partial pressure
analyzers, and gas chromatographs. The methods used with this equipment are listed in Table 3. Methods that can be used
with both vacuum and pressure systems are discussed in detail in the section "Leak Testing of Pressure Systems Using
Specific-Gas Detectors" in this article.
Manometers. Testing for leaks in systems under vacuum can be done by observing the manometers installed on the
system. Although manometers are not as adaptable to locating a leak as regular leak detectors, some indication of leakage
can be attained. Because most vacuum gages will react or change their reading in the presence of a tracer gas, spraying
tracer gases over individual portions of the vacuum system will produce a gage response and will provide some indication
of leak location. The location can also be roughly established by isolating various portions of the system and observing
the gage response. The primary evidence of a leak in the system is failure of the system to achieve anticipated pressure
levels in a certain pumping time as based on past or calculated performance. However, consideration must be given to
other gas loads (virtual leaks) that may exist in the system. Sources of these gas loads are system contaminants, such as
water, outgassing of chamber surfaces and test items, and outgassing of entrapped areas such as unvented O-ring grooves.
Pressure-change gages having a pressure-change sensitivity of 0.13 mPa (10
-6
torr) are available for vacuum applications.
Ionization gages will measure vacuums to 1.3 pPa (10
-14
torr) and can be used to measure changes in pressure. The
sensitivity of all pressure-change methods is time dependent (that is, the mass changes per unit of time); therefore, the
longer the test duration, the more sensitive the method.
Halogen Detectors. Heated anode halogen detectors can also be used on vacuum systems. With special adapters, they
can be used down to 0.13 Pa (10
-3
torr). Leaks are detected by using an externally applied tracer gas.
Mass spectrometers are well suited for use as leak detectors in vacuum systems because the detector tube of a mass
spectrometer must be maintained at a vacuum. Also, mass spectrometers usually have built-in vacuum-pumping systems
that can be directly coupled to the system being tested. Helium mass spectrometers are most commonly used. Leaks are
detected by spraying or enveloping the outer surface of the system with helium.
Ionization Gages. In addition to detecting leaks in vacuum systems by pressure-change instruments, ionization gages
can be used to detect the presence of specific tracer gases. Because each gas traveling through the gage ionizes
differently, equal flows of different gases will produce different readings. Leak rates for molecular or laminar flow can be
related to ionization-gage readings in response to a tracer gas through:
(Eq 9)
(Eq 10)
where Q
a
is the flow of air (in torr liters per second), ΔR is the change in ionization-gage reading (in torr), S
a
is the
pumping speed (in liters per second), is the gage sensitivity factor, v = η
a
/η
t
(where η
a
is the viscosity of air and η
t
is
viscosity of tracer gas), and μ= (M
a
/M
t
) (where M
a
is the molecular weight of air and M
t
is the molecular weight of
tracer gas). Values of σ for ionization gages, and of the gas factors (σ - 1) and [(σv/μ) - 1], for several gases are as
follows: