Problems With Corrosive/Reactive Gases and Contamination
First let’s establish some definitions for discussion purposes:
A material which is either corrosive to metal or can produce a corrosive material when combined with another material. An example of corrosive gases would be HCl, or HF, which are directly corrosive, or Chlorine which can produce HCl when exposed to water vapor.
A gas that will easily react with another substance to produce something undesirable to the operation of the flow controller. Now H2 is a very reactive gas, but it is most likely to react with O2 producing water, which is not likely to happen in most flow controller applications but it would certainly mess up the flow controller. Silane on the other hand can react with moisture in the air (via a system leak) to produce SiO2 (sand) among other things that can cause really bad things to happen to a flow controller. It also can produce a fire and explosion when exposed to air in large quantities but that is another discussion altogether.
Either of these requires the help of some type of additional material, like moisture, which is usually introduced by leaks in the gas system, especially sub-atmospheric systems, or through breaches of the system integrity. These breaches can be caused by improper maintenance procedures, faulty components, or gas tank changes which are not performed with adherence to proper purge procedures.
One result of these actions will be the production of some form of a contaminant inside the flow controller. Let’s look at what the impact of that contamination would be.
Contamination is the presence of any substance, other than the intended process gas, that adversely affects the operation of the device. This contamination can be particulate, such as rust, dirt, or precipitation from a contaminated process gas. It could also be the presence of any substance that affects the proper operation/calibration of the device, such as another process gas from a faulty isolation/check valve, moisture, even the presence of condensed (liquid) process gas (BCl3 is a good example of a very bad gas).
There are several ways for contamination to affect the flow controller operation.
The sensor tube system produces an electrical signal indicating the flow through the device by sensing changes in heat transfer proportional to the mass flow of the desired process gas though the sensor tube. This requires the precise characterization of the heat transfer for the expected process gas at different flows. The introduction of a material with different thermal characteristics from the calibrated gas in the sensor tube can cause a dramatic change in the predicted heat transfer characteristics for the calibrated gas, or worst case a complete blockage. This will result in a calibration offset and/or erratic control. In this case it does not matter what the contamination is, as long as it results in changing the conditions that were present during the initial calibration. That means the sensor could be contaminated during the initial calibration and the ensuing operation of the device would be unaffected as long as it remained constant. Unfortunately, contamination is rarely uniform nor remains in a steady state for very long.
This type contamination is the most critical as the sensor tube can be as small as .005” in diameter, and it takes very little material to change the thermal characteristics of the tube. The low volume and the relatively low flow velocity through the sensor tube makes it very difficult to successfully purge particulate contamination. Even upon disassembly this contamination is not easily observed and may require a liquid purge, a mechanical cleaning of the tube, or a complete replacement of the sensor to correct.
The bypass creates backpressure on the gas stream, thereby forcing a ratio of the gas flow through the sensor tube that is proportional to the total gas flow through the device. Any contamination of the bypass that results in a change of the flow characteristics will change the relative flow through the sensor tube resulting in a calibration offset, a control problem, or worst case, a complete blockage. This type contamination is usually very visible and easily corrected but requires disassembly and recalibration of the device.
The valve controls the flow of gas to a desired set point. Contamination of the valve can cause mechanical binding resulting in erratic or no control, or incomplete closure which can cause leakage at zero set point or erratic control at lower set points. This contamination will usually be very visible and easily removed but requires complete disassembly and recalibration of the device.
This means the most likely, initial, negative impact of corrosive/reactive gases on the flow controller will be the direct result of contamination of the device.
Sources of Contamination
External corrosion is the result of a chemical degradation of the external components of the device, and is usually caused by a leak of an acid (HCl) or acid forming gas (like Cl) within the piping system/gas panel. The ensuing corrosion will be seen (or even smelled) first on the non-stainless steel external surfaces of the gas panel, around the electrical connector, the bolts securing the connector or cover, or the mounting brackets. As many gas panels are fitted with an exhaust system to remove these fumes it may be very difficult to spot the affect immediately. Unless the flow controller is the source of the leak (and this is very rare indeed), it is likely it will function normally until a complete structural failure occurs. This would most likely be the failure of the electrical connection, or in some cases a failure of the bolts that hold the device together. Needless to say this is not a flow controller contamination problem per se, but it is a problem we see all too frequently.
Internal Corrosion is the result of a chemical degradation of the wetted metal surfaces of the device. In most process applications this would be the result of the presence of an acidic process gas (like HCl or HF), or the formation of an acid by the reaction of the process gas (like Cl) with some type moisture introduced into the system. The device is literally rusting inside.
Although the effects of internal corrosion can result in a catastrophic failure of the device, such as a breach of the sensor tube, it is much more likely the effect will be seen as a functional problem resulting from the particulate generated by the corrosion process. The effect is essentially the same as contamination and in most cases the device won’t know the difference, especially in the early stages of corrosion. Severe internal corrosion is usually not observed on a device when in use, as the functionality of the device will be affected long before the corrosion can advance. Severe internal corrosion will most likely be observed on a device that was removed from a system without proper purging, and exposed to ambient conditions for a period of time before disassembly.
Post Removal Contamination/Corrosion
No amount of purging can guarantee the device, especially the sensor tube, will be fully purged of residual process gas. This is especially true for corrosive gases with low vapor pressure (like BCl3) that can condense into a liquid inside the sensor tube or reactive gases like Silane that can deposit precipitates under the presence of moisture from the ambient air. The gas flow and velocity that can obtained through the sensor tube when purging the device under normal purge conditions will not be sufficient to completely remove the contamination nor keep it from continuing to corrode/react after the device has been removed from service. It is very important when servicing corrosive/reactive gas devices to get them serviced as soon as possible after removal from the system to reduce the amount of continuing corrosion or contamination. Always be suspect of any device that has been previously installed in a corrosive/reactive system regardless of the purge procedures used before removal. If there is ANY visible sign of contamination (white or brown powder, green discoloration, visible particles) on the inside of the inlet or outlet fitting there is a good probability the device is contaminated much more severely internally. This may not render the device completely inoperable but it will most likely result in some form of functional or process related problem in the future.
The Gas Delivery System
The gas delivery system itself can be a source of contamination of the flow controller. This can be from the effect of corrosion of the metal components of the system, or generation of particulates through chemical reaction of the process gas. Either of these will ultimately produce the same end result, stuff gets trapped in the flow controller. The bypass and sensor tube of the flow controller can be the most restrictive path in the gas system except for the gas filters, if filters are installed. This means any particulate generated upstream of the device and downstream of the filter will be captured in the device. It is highly unlikely this contamination will pass through a flow controller without affecting the operation of the device in some way. Even with a filter in place some contamination can reach the flow controller, especially if it is the result of process gas reaction as this reaction can continue to take place after passing through the filter.
Most high purity, metal seal, flow controllers on the market today have an internal surface finish that is far superior to the finish found inside most of the other components of the gas delivery system. This means the surface area has been polished and chemically treated to make it very resistant to corrosion. The presence of contamination in the flow controller as the result of corrosion is far more likely to be the result of system component corrosion that has been captured by the restrictive gas path of the device than from the device itself. Corrosion of the internal surfaces of the flow controller can be confirmed by the observation of pitting or discoloration of the surfaces after removal of the residue. This means that the flow controller can usually be cleaned and returned to service with no effect on the integrity of the internal surfaces if the gas delivery system was the source of the contamination.
This also means the replacement flow controller in the corrosive application has a very high probability to exhibit further performance problems as this corrosion continues to produce particulate, even if the source of the corrosion has been addressed, as it is very difficult to completely clean an entire gas system. This is especially true of reactive gases like Silane that can produce a fine powder throughout the gas system when exposed to even small quantities of moisture from a leak or improper tank change. This will cause an almost instantaneous contamination problem that will be seen as an operational problem with the flow controller that has now begun to accumulate particulates. The flow controller can now be cleaned and returned to service, or replaced by another, but the residual contamination is now deposited throughout the system. It may even continue to be produced if the problem is still present. The replacement flow controller may show signs of residual contamination immediately upon installation or very soon thereafter. This can develop into a continuing spiral of flow controller change outs that will continue until all the particulate matter has either been captured by the unlucky flow controllers or purged from the system.
We have had many comments from customers who say they never had trouble with their flow controllers on a system until they started having trouble; then, they seem to have sporadic problems thereafter. In this case lack of completely removing the system contamination would be the most likely culprit.
If we assume the flow controller has been properly configured for the operating parameters of the gas system in which it is installed (and you can bet that is covered in another tutorial section), contamination is by far the most common cause of flow controller problems we encounter. It can also be the most frustrating to identify and correct, as the effect may not show up for days or weeks after an event that triggers corrosion or contamination. Even worse, it may be the cumulative result of several small events that show no immediate effect alone. Perhaps the most frustrating part is the ongoing problems that can continue after replacing the “bad” flow controller.
My old process guru once said, “There are two types of process engineers using Silane gas. Those who have contaminated the system and those who are going to.”
This could be applied to most systems using corrosive or reactive gases.