Leak Detection and Troubleshooting 7.1 Leak Detection An ideal vacuum chamber should maintain for ever the vacuum reached at the moment of its separation from the pumps. But a real chamber presents a rise in pressure after being isolated from the pump. Pressure rise can be either due to leak or due to outgassing or due to both. Leak detection is perhaps one of the most important but some times one of the most tedious and frustrating aspects of vacuum technology. However much of this unnecessary effort can be avoided if a correct approach is made to the original design and construction of the vacuum system. In practice there are really two main aspects of leak detection, namely testing components and testing completed systems and any particular technique may be more or less appropriate to the one than the other. However it is first necessary to look at the nature of the leaks that can occur in all-vacuum systems. 7.2 Real and virtual leaks: Leaks are normally referred to as real or virtual, where in the former the gas passes from the external atmosphere into the vacuum chamber and the latter arises either due to the evolution of gases or vapors trapped inside the vacuum envelope in holes or channels or due to desorption of adsorbed molecules or vapors on the inside walls and components in the vacuum system. In the latter case the most important example of this is adsorbed water vapor. The leak rate Q L can be defined as the quantity of gas, which enters the vacuum space per unit time from real or virtual leaks or both. …(7.1) Where dp/dt is the rate of rise of pressure in a closed volume, v, isolated from the pumps. If the system is being continuously pumped the rate of removal of the gas is equal 133
Transcript
Leak Detection and Troubleshooting
7.1 Leak DetectionAn ideal vacuum chamber should maintain for ever the vacuum reached at the moment of its separation from the pumps. But a real chamber presents a rise in pressure after being isolated from the pump. Pressure rise can be either due to leak or due to outgassing or due to both. Leak detection is perhaps one of the most important but some times one of the most tedious and frustrating aspects of vacuum technology. However much of this unnecessary effort can be avoided if a correct approach is made to the original design and construction of the vacuum system. In practice there are really two main aspects of leak detection, namely testing components and testing completed systems and any particular technique may be more or less appropriate to the one than the other. However it is first necessary to look at the nature of the leaks that can occur in all-vacuum systems.
7.2 Real and virtual leaks:
Leaks are normally referred to as real or virtual, where in the former the gas passes from the external atmosphere into the vacuum chamber and the latter arises either due to the evolution of gases or vapors trapped inside the vacuum envelope in holes or channels or due to desorption of adsorbed molecules or vapors on the inside walls and components in the vacuum system. In the latter case the most important example of this is adsorbed water vapor.
The leak rate QL can be defined as the quantity of gas, which enters the vacuum space per unit time from real or virtual leaks or both. …(7.1)Where dp/dt is the rate of rise of pressure in a closed volume, v, isolated from the pumps. If the system is being continuously pumped the rate of removal of the gas is equal to the rate of entry of the gas through the leak. Then we can write,
…(7.2)Where S is the pumping speed and Pu is the ultimate pressure in the system. Hence it is meaningless to speak about a 'large leak' unless we are able to specify the volume of the system. These considerations determine whether a leak is significant or not and in a continuously pumped system a leak may be present, which is small enough to be ignored. For example, in chamber with volume 125 litres and surface area 1.5104 cm2, the total surface outgassing rate (for stainless steel) would be 1.510-8 mbar l s. With a pump of speed 1000 l s-1, the pressure with out leaks would be 1.510-11 mbar. If we now have leaks of various sizes, the resultant pressure is shown in Table 7.1 below.
Table 7.1:
133
QL (mbar l s-1) QL (mbar l s-1) P (mbar)1.010-5 1.0001510-5 1.01510-8
1.010-9 1.610-8 1.610-11
1.010-10 1.5110-8 1.5110-11
From these figures it can be seen that a small leak, well with in the detection limits of all teh helium leak detectors, has little effect on the final pressure. The decision has to be made whether a leak can be ignored.
It should also be emphasized that the leak rate will be independent of the chamber pressure provided that the pressure is less than about 1 mbar, because the pressure difference will be constant and equal to 1 atm. The rise of pressure with a constant real leak into fixed volume is shown as a function of time in Figure 7.1 (a) and the straight line indicates that the gradient of the line, dp/dt, is constant.
However if a similar graph is plotted for a virtual leak then the initial high rate of rise of pressure gradually falls off to zero as shown in Figure 7.1 (b), when the rates of adsorption and desorption of the gases or vapors from the surfaces inside the vacuum chamber are in equilibrium. The familiar case when a real leak and the virtual leak are both present is shown in Figure 7.1 (c) Hence the common practice of using the pressure versus time graph to identify a real leak must be used with caution.
Methods of leak detection:
There are essentially two methods for the detection of leaks, namely when an excess pressure of the search gas is maintained outside (Tracer probe technique-Method O) or inside (Detector probe technique-Method I) the system under test, using a number of different types of leak detector in each case.
In Tracer Probe method (fig7.2a), a stream of test gas is spread on the suspected area and the gas penetrating into the system is pumped via the detector.
(a)
134Test gas Container
Probe 2Detecto
r1Pump
System3
5
Fig. 7.1 Pressure versus time curves if (a) a real leak, (b) a virtual leak or (c) both are present
(b)
Fig 7.2: (a) Tracer probe and (b) detector probe techniques.
In detector probe method (Fig 7.2b), the test gas is filled into the system under test and a detector probe (sniffer) connected to the leak detector is passed over the suspected area to receive the test gas. He gas is usually used as test gas because it flows through fine leaks (due to small atomic size) and background error is negligible (since partial pressure of He in atmosphere is very low).
7.3 Leak detectors
The mass spectrometer is the most sensitive and most important leak detector but many small companies and institutions cannot justify the expense of one of these instruments. Hence a number of alternative simple method. Leak detection is also included in this section.
In each case it will be indicated whether it is suitable for method (I-Fig 7.2b) and or (O-Fig 7.2b).
Soap bubbles (I): In this case the probe gas will normally be compressed air but greater sensitivities can be obtained using helium. The soap solution can be brushed on the suspected areas and bubbles will be seen if a leak is present, However it is more satisfactory if the chamber can be totally immersed in the liquid and the bubbles can be seen rising to the surface. It is possible to obtain higher sensitivities using liquids of lower surface tension.
Leak covering (O):A suspected leak can be covered over with a tape or low pressure 'plasticine' such as Apeizon Q. If the leak is present then a fall of pressure will be observed. However this method should be used with care because when the tape or plasticine is removed the leak may be temporarily sealed but will almost certainly reappear at a later stage. This method should not be used for UHV applications.
135
Flows through fine leaks. Partial Pressure is negligible in atmosphere.
Test gas (He)Containorr
Probe
5
System 3 Detector
2
1Pump
Thermal conductivity detector (I and O): As the thermal conductivity varies for different gases then a Pirani or a thermocouple gauge can be used as a leak detector. One of these devices is included in most vacuum systems and a leak can be indicated by an apparent change in the pressure when another gas e.g. carbon dioxide is directed over the leak. Low priced hand held monitors using the thermal conductivity principle are widely available and can also be used to detect the escape of domestic and industrial combustible gas from meters, valves and pipeline joints. Leak detection over welds, joints, seals etc can be carried out using helium or carbon dioxide as the trace gas.
Ionization gauge and ion pump detectors (O): Most high vacuum systems include some type of ionization gauge. The ion gauge current and the ion pump current are both dependent on the gas species in the system or pump and changes can be seen when a leak is present and covered with a search gas as used for the thermal conductivity gauge.
Useful safe gases for both methods are helium, carbon dioxide and argon. Acetone and isopropyl alcohol can be used with care.
Mass spectrometer Leak Detector (I and O): The mass spectrometer leak detector is an essential piece of equipment in almost every laboratory involving vacuum equipment and the many associated technologies. This may be an already existing residual gas analyzer (RGA), whether it be magnetic sector or quadrupole type, but it does not need to have a very high resolution because it is only necessary to separate the two most commonly used gases, namely hydrogen and helium at mass numbers 2 and 4 respectively. Nevertheless the dedicated mass spectrometer leak detector is usually of the magnetic sector type because of its higher resolution at lower mass numbs and its higher sensitivity compared with the quadrupole.
Helium is usually chosen as the search gas because the molecule is small and inert, it has a higher particle velocity than any other gas at any particular temperature except hydrogen, has no environmental hazards, is non-corrosive ad perhaps most importantly is unlikely to be present in the vacuum system for any other reasons. In exceptional cases hydrogen may be preferred because of its higher sensitivity.
The operation of a spectrometer consists of four main steps: (i) ions are cerated from the gas phase usually by electeron impact, (ii) the ions are accelerated to known kinetic energies in a chosen direction, (iii) ions entering the analyzer are subjected to an arrangement of electric and/or magnetic fields which separate them on the basis of their mass-to-charge ratios, and (iv) the separate ions are detected upon arrival at a collector. The basic objective is to provide usually an ion current signal or change in an ion currewnt related to the number density of the species. In principle any combination of the various types of ion sources, analyzers anddetectors which form mass-spectometer system can be used with any probe gas (e.g. hydrogen, helium or argon) for leak detection.Fig.7.3 gives a typical configuration of a mass spectrometer. Positive ions produced in the ionization chamber are accelerated by a potential V. Then, these ions are allowed to
136
pass through a magnetic field where they are sorted out according to their values of e/m, where m is the mass of the ion and e is the charge. The radius path of any given type of singly-charged ion is determined by the mass of the ions, their charges and the ratio of the electric and magnetic fields. The mass-spectometer is tuned to detect usually helium mass(4 amu) and hence V and H are fixed. The system under test is connected at the inlet to the mass spectometer. Helium is sprayed externally onto the system and when the helium gas comes over a leak, it passes through the leak into the mass-spectometer. The ion current recorded is a measure of the size of the leak. In the vacuum test method, with helium as the probe gas, sensitives of the order of 10-9-10-10 Torr-litre/sec are obtained at total pressure up to 10-4 Torr. The sensitivity decreases with increase in test pressure.
The conventional spectrometer leak detector is normally a portable unit containing its own vacuum system including a rotary and a diffusion pump, liquid nitrogen trap, gauges, valves and all the associated instrumentation and electronics as shown schematically in Figure 7.3. A Penning gauge is normally used because of the continual cycling to atmospheric pressure.
When the vacuum system is in operation the component or system under test is connected to the 'test port' with valves b and d closed and valves a and c open. A number of helium leak detectors have a second rotary ump for roughing out the system under test while the diffusion pump is kept under working conditions by its dedicated pump. When the pressure is low enough valve a is closed and valves b and d are opened. The mass spectrometer filament is switched on when the pressure on the Penning gauge is less than 510-4 mbar and the system under test is probed with the helium search gas. If a leak is detected this is indicated by an audio signal and /or an analogue meter on the electronic panel. The sensitivity of the detector can be increases by (i) increasing the gain on the electronic amplifier and/or (ii) reducing the effective speed of the diffusion pump by adjusting the speed control valve but still maintaining the minimum pressure to operate the filament.With this arrangement the minimum detectable leak is about 10-12 mbar l s-1. For calibrating the leak detector, one can use standard leak with attached pure He cylinder (small), commercially available. Nevertheless, it should be emphasized that leak testing is somewhat of an acquired skill and thus the varies 'tricks of the trade' can only come with experience.
A further type of detector which requires no liquid nitrogen is shown in fig 7.4 and uses a reflex or contra-flow type of leak detector.
137
Fig. 7.3 Schematic diagram of a conventional mass spectrometer leak detector
The mass spectrometer is kept clean and at constant pressure and the helium back-streams through a special diffusion pump where the rate is variable, controlled by the heater power. A dual port entry turbo pump or molecular pump can also be used.
Of course many of the new generation of leak detectors reflect the influence of modem technology and computer control. For example, many use microprocessor control with fully automatic start-up and built-in tests for such things as the device sensitivity and calibration using a built-in standard leak. In other cases there are software supported functions for fault finding in the leak detector vacuum system and many have an automatic inlet which adjusts the opening to the test port.
However if a leak appears to have developed in a vacuum system it is nearly always helpful to answer these ten questions before taking any further action.
(1) Has the pressure in the system changed since yesterday?(2) Has a different type of gas been admitted into the system?(3) Can any sections of the vacuum system be isolated?(4) Has the gauge calibration been checked?(5) Are there any vapours present?(6) Has a real leak been established even though not found?(7) Can you tolerate the leak that has been established?(8) Have you used some different type of cleaning material after the last exposure
to the atmosphere?(9) Have you added to or modified the interior of the vacuum system?(10) Is there any evidence of an air leak from an RGA spectrum with peaks
corresponding to masses 14, 28 and 32?
138
Mass spectrometer
Diffusion or turbo pump
Backing Roughing
Test port
Fig. 7.4 Reflex or contra-flow type of leak detector
Although more than one option is available for leak detection, a strategy needs to be planned to optimize the effort. A job needing a leak test must examined with regard to its end-use, range of operation parameter, materials used in its construction and dimensions, before a procedure for leak test can be worked out.
Various physical processes that takes place in vacuum between gas molecules and surfaces, such as adsorption and desorption, permeation and diffusion, sublimation and condensation, can also interfere with the process of leak detection. And hence, for a better interpretative use of leak- detection parameters, the above contribution needs to be taken into account. Leak detection, to be reliable, required a lot of patience on the part of the operation, who besides must understand that for every type of leak detector there is a methodology to be followed. Primarily, the nature of the probe gas – its size, molecular weight, to be followed. Hence, its is obvious that the methodology has to be planned, starting right from selection of the materials of construction of vacuum chamber, to joining methods, pumps and gauges and their locations.
7.4 Acceptable Air Leakage Rates in process industries (Rough Vacuum):
Assumption that the acceptable air leakage rate correlates directly with the volume of the vessel is wrong. Because of leaks across system components, vessel B in Fig, 7.5 requires more maintenance to hold air leakage at an acceptable level than does vessel A, a vessel of equal volume.
Air Leakage Calculations:
As already noted, air leakage rates into rough-vacuum system are directly related to the quality of vessel maintenance. Mass transfer attributed to leakage occurs by pneumatic flow through leakage paths that exit as a result of errors in fabrication, assembly, and installation of the vessel.
1397.5 Using vessel size to estimate acceptable air leakage can result in serious
errors
Table 7.2 Specific Leak Rates for Rough-Vacuum System Components
Rotary seals: Packing glands Mechanical seals Valves used to isolate system: Plug cock Ball Globe Gate Valves used to throttle control gas into vacuum system Access ports Viewing windows
The following procedure for estimating acceptable leakage rates in to a vacuum system will permit the designer to write responsible specifications for rough vacuum equipment:
1. Estimate acceptable air leakage resulting from metal porosities and cracks along weld lines from the following questions:
0.1 P < 1 torr W = 0.026 P0.64 V0.60 …………….(7.3)
. 1 P < 10 torr W = 0.026 P0.34 V0.60 -------------(74)10 P < 100 torr W = 0.032 P0.26 V0.60 -------------(7.5)100 P < 760 torr W = 0.106 V0.60 -------------(7.6)
140
Where P = System operating pressure, torrV = volume, ft3
W = air leakage resulting from metal porosities and cracks along weld lines, lb/hr
2. Estimate acceptable air leakage stemming from leakage around static or rotary seal, Valves, access ports, and other features required for process operation from Eqs. 8.7 to 8.10 and the specific leak rates presented in table 8.2(w 10lb/hr, the maximum acceptable air leakage for a component):
0.1 P < 1 torr w = P0.64 ……………………(7.7)1 P < 10 torr w = P0.34 -------------(7.8)10 P < 100 torr w = 1.20 P0.26 -------------(7.9)100 P < 760 torr w = 3.98 -------------(7.10)
Where D = Scaled diameter, in W = acceptable air leakage rate assigned to a system component. Lb/hr
= Specific leakage rate, lb/(hr)(in)
3. Calculate the total acceptable air leakage rate WT, lb/hr, by adding W to the sum of
the leak rates assigned to the individual system components Σ w.WT = W + Σ w --------- (7.11)
In estimating an acceptable air leakage rate, do not overlook the implications of process contamination by air leakage if leakage in to the vacuum system could result in termination of a reaction, serious oxidation of the process materials, or in any way adversely affect the process or present a safety hazard, the acceptable air leakage rate may be considerably less than that predicted by the procedure presented here.
Leak Rate Measurement:
1. Evaluate the system to a final pressure in the range 10 to 100 torr.2. Close the isolation valve between the vacuum vessel and the vacuum pump and
shut the vacuum pump down.3. Recorded the time required for the pressure in the system to rise from P1 to a
higher pressure P2 never permitting the pressure to rise above 0.53 times atmospheric pressure .
4. Calculating the air leakage rat from the equation
Leakage, lb/hr = …………..(7.12)
Where V = total system volume , ft3
P = pressure, torr t = time, min T = temperature, 0R
141
The vacuum vessel can be isolated from the vacuum pump by closing a valve. An isolation valve is not normally installed, for example, between a polymers reactor and the vacuum pump. Polymer entrained in the vapor steam will accumulate in the valve and plug the line. If the steam jets normally used in this service exhaust to a barometrically sealed after condenser, a drop test can still be used to determine the air leakage rate.
Steam flow to the jet (and cooling water flow to direct contact condensers) is shut off following evacuation of the vacuum vessel. The result for the drop tests under these conditions are less accurate but never the less useful, because results for consecutive tests provide a basis for assessing the change in leak rate over time. The initial change pressure with respect to time will be nonlinear if the pressure in the system approaches the vapor pressure water at ambient temperatures. Water vapor backstreaming from the barometric seal causes an increase in total pressure. Leakage past the valves used to shut of steam flow to the jet will also affect results.
Temperature has no significant effect on air leakage rates the drop test is normally conducted at ambient temperatures.
