Oxygen Cleaning: A Validated Process Is Critical For Safety

David Escobar, Director of Engineering at Metso Automation

Industrial oxygen is used for many purposes: in a basic oxygen furnace for making steel, water pollution countermeasures, including sewage treatment, habitability and superfund site rehabilitation, and chemical processes such as production of vinyl chloride, nitric acid, epoxyethane and hydrogen peroxide. It is also used for medical treatment, life support in harsh environments and industrial gasses for welding and other processes.

The production of oxygen has risen from approximately 470 billion cubic feet in 1991 to over 1.5 trillion cubic feet in the U.S. and more than 4 trillion cubic feet globally in 2014.

Oxygen is the most common oxidizing gas and is, of course, highly reactive. When dealing with an oxygen-enriched environment, it is important to control the sources of ignition. Ignition can be caused by many things, among them:

  1. Electrical arcs, which can come from electrical equipment or even static discharge
  2. Friction, which can be generated by the sliding contact of materials within the oxygen-enriched environment
  3. Impact of particles or projectiles internal or external to the enriched environment can generate heat
  4. Resonance, which is vibration-induced heating
  5. Heat of compression (HoC) is the most common cause of explosion due to contamination. Heating is caused by the adiabatic compression of a fluid; this is often called auto-ignition.

Auto-ignition is the phenomenon of spontaneous ignition of a fuel source due to the heat generated by the sudden compression of a gas or HoC. When a valve in a high-pressure or high-velocity oxygen flow is opened or closed quickly, the kinetic energy is converted to increased temperature and potential energy in the form of increased pressure. If the temperature generated by the compression exceeds the temperature needed to ignite non-metallic seals or even the pipe itself, the result is a spontaneously explosion or auto-ignition. When this happens in oxygen systems, the effect can be devastating.

A fire in a process plant

Because the HoC is substantial and can generate thousands of degrees of temperature even at moderate pressure ratios, oxygen systems are designed to limit the pressure drops to control HoC and limit temperature within the autoignition temperatures of the system components.

Thus, it is absolutely essential that contaminants, which can introduce lower auto-ignition temperatures than even the non-metallic seats and seals, be removed from any oxygen system. Any method that achieves the desired cleanliness level is acceptable. CGA 4.4 and the recently issued MSS-SP-138 provide excellent recommendations for cleaning processes.

Oxygen Cleaning A Validated Process is Critical for Safety 2
A technician moves hardware in a clean room using proper protective equipment


Basically, anything that promotes combustion or impact product purity is considered a contaminant. ASTM G93 categorizes contaminants into three types:


  • Volatile Organic Compounds (VOC)
  • Hydrocarbon-based greases and oils


  • Nitrates
  • Phosphates
  • Water-based detergents and cutting oils
  • Acids/solvents


  • Particles, lint and fibers
  • Dust – Weld slag, etc.

Specifications vary on cleanliness level, methods and validation, and include how much residue is acceptable, what method of cleaning can be used and what kind of inspection must be conducted.


Mechanical cleaning is used to remove scale, coatings, paint, weld slag and other solid contaminants and can include grit or ice blasting, wire brushing and grinding.

Aqueous cleaning can be with hot water and steam cleaning or alkaline cleaning. Hot water and steam cleaning is effective against water-soluble contaminants, and is normally used with detergent. Alkaline cleaning uses caustic salt in water to create a highly alkaline solution. It is effective against hydrocarbon oils, grease and waxes, and generally is enhanced by agitation and/or jet spraying. Typically this is used for industrial parts washers. This process is greatly enhanced by ultrasonic agitation, but the solvent residue must be removed as well.

Semi-aqueous cleaning uses hydrocarbon solvent and water emulsion, which is effective for removing heavy contaminants from parts like heavy grease wax or hard to remove soils. Emulsion may require agitation to maintain the mixture, and parts must be rinsed before the emulsion can dry. Otherwise, contaminants may re-deposit on the part that was cleaned.

Acid cleaning varies substantially based on the acid used.

  • Hydrochloric acid is used to remove scale, rust and oxides. and to strip platings (chrome, zinc, cadmium, etc.) and other coatings
  • Chromic and nitric acid are used to for passivating, deoxidizing, brightening and removing alkaline residues in addition to cutting oils
  • Phosphoric removes oxides, light rust and fluxes

Acids must be removed completely from the part prior to drying and, depending on the acid strength, may need a neutralizing process.

Solvents can be used without water dilution or emulsion. Alcohol is a common solvent often used to revisit areas of concern identified by black (UV) light inspection. Solvents like alcohol evaporate completely, leaving no residue.

Vapor degreasing is a process in which a solvent is heated until it vaporizes, while the part is maintained at a lower temperature. The solvent then condenses and dissolves contaminants. The part must be oriented so that the condensed solvent can drain from the part by gravity. This method is very effective for inaccessible areas on parts but requires a contained environment for the part during the process.

Any combination of cleaning methods that achieve the desired cleanliness level is acceptable.


Visual inspection can be direct, including white light, which is effective in detecting contamination down to 500 mg per square meter. UV (black) light visual inspection identifies contaminants that fluoresce and is effective in detecting contamination down to 40 mg per square meter.

Indirect visual inspection is done in two ways: wipe test and solvent filtering. A wipe test can identify contaminants in locations that have no direct line of sight. Typically, both white light and UV light are used on the wipe cloth, and are effective in detecting contamination down to 30 mg per square meter. Solvent filtering rinses the inaccessible area in solvent, which is then filtered to capture contaminants. The filter is then visually inspected and can detect 100 ml per square foot of low residue solvent and it also uses white and UV light.

Oxygen Cleaning A Validated Process is Critical for Safety 3
White light inspection of cleaned surfaces

Quantitative inspection is done by evaporating the solvent used for cleaning and obtaining the weight of the remaining effluent. Acceptable levels of residue vary according to user requirements.


Clean room: This provides a designated location where the environment limits dust airborne particles, where clean tools and clean assembly and test equipment can be stored. It can also provide controlled lighting for visual inspections.

Clean test equipment: Pressure test equipment contains contaminants in hoses and pumps. If a test machine cannot be dedicated for clean testing, give special consideration to cleaning of test equipment or alternate testing with clean gas.

Packaging: After cleaning, give specific instructions on how to package the product to preserve cleanliness in shipping and subsequent storage. Consider the role of desiccant as a possible contaminant. Use compatible products or control desiccant to prevent contamination. Consider the addition of actuation and accessories to the valve. Can the actuator be installed and set up without violating the protection? If the protection is compromised, are there procedural steps to identify and remediate any contamination?


Oxygen cleaning is used to remove contaminants that can significantly reduce the temperature of auto-ignition. There are many methods for doing the actual cleaning. CGA 4.4 and the recently issued MSS-SP-138 provide excellent recommendations, but any method that achieves the desired cleanliness level is acceptable. It is important to know what level of cleanliness your standard process produces. Process validation using a quantitative measurement allows the supplier to have confidence in process quality when using qualitative inspections for production work.

Editor’s Note: The following article was posted on April 20, 2015 in Valve Magazine.com. It is reprinted here for the Reaction Research Society (RRS) with permission from the author and Valve Magazine. The information here is very useful in amateur rocketry and is intended to make our readers aware of the importance of a proper oxygen cleaning process for lines and valves. High purity oxidizers must be handled with care and cleanliness is of paramount importance. The RRS would like to thank David Escobar of Metso Automation and Judy Tibbs, Director of Education at the Valve Manufacturers Association and Editor in Chief of VALVE Magazine.

David Escobar is director of engineering at Metso Automation. Reach him at david.escobar@metso.com.

CGA refers to the Compressed Gas Association. Founded in 1913, the CGA is an organization dedicated to the development and promotion of safety standards in the industrial, medical and food industry. The CGA is comprised of over 110 member companies worldwide working together through the committee system to create technical specifications, safety standards and educational materials; to cooperate with governmental agencies in formulating responsible regulations and standards; and to promote compliance with these regulations and standards in the workplace.

For more information, go to the CGA website:


MSS refers to the Manufacturers Standardization Society of the Valve and Fittings Industry. Standard practices (SP) documents are available related to many applications including the standardized practice of oxygen cleaning (ANSI/MSS SP-138). ANSI or the American National Standards Institute has adopted the standard for oxygen cleaning of valves and fittings.


ASTM stands for the American Society for Testing and Materials. It is now an international organization known simply as “ASTM International” with its headquarters in West Conshohocken, Pennsylvania. ASTM publishes voluntary consensus technical standards including ASTM G-93 for the Standard Practice for Cleaning Methods and Cleanliness Levels for Material and Equipment Used in Oxygen-Enriched Environments.

For more information, go to the ASTM International website:


April 2019 meeting

The RRS held our monthly meeting on April 12, 2019 at the Ken Nakaoka Community Center in Gardena. We had a full agenda with the 2019 RRS symposium just around the corner on Saturday, April 27th.

The symposium is just around the corner

We first welcomed two new members, Keith Yoerg and Jonathan Martinez. Keith is active with Tomorrow’s Aeronautical Museum (TAM) at the Compton Airport and has given many educational programs to local schools. He’s also a graduate of USC and a former member of their Rocket Propulsion Laboratory (RPL). Jonathan Martinez joins the RRS as a student member from Compton High School. He’s been working at TAM and the RRS hopes to help him in his new project to hot-fire a liquid rocket.

Keith Yoerg (left) watches Waldo Stakes (right) show off the gas generator injector he brought to the meeting.
New RRS member, Jonathan Martinez (left) and Wilbur Owens (right) at the April 2019 meeting of the RRS.

We next talked about the recent launch event with LAPD CSP and Compton Elementary. The “Rockets in the Projects” program is going strong and we were glad to welcome Compton Elementary to our workspace and launchpad in the Mojave Desert.

Dave Crisalli, our pyro-op for the event, gives the safety briefing to all attendees including Compton Elementary, LAPD and USC RPL

Under very pleasant weather, we had a good launch event starting with a tour, safety briefing and the kids finally getting a chance to see their rockets fly into the blue sky. Osvaldo had a seventh alpha rocket with a parachute system, but somehow failed to deploy. USC static-fired a six-inch custom solid motor.

An alpha assembled at Compton Elementary streaks away from the box rails at the RRS MTA
USC’s six-inch solid composite grain motor burns for full duration at the RRS MTA. A second motor will be integrated into the Traveller IV vehicle that USC will launch from Spaceport America in New Mexico

After Compton Elementary and LAPD CSP went home, Osvaldo, Frank, Larry and I did a little reconnaissance for the alphas we flew at the event. We were able to find 3 of the original 6 and one more alpha from the past MTA launch event. The higher level winds have been carrying the alphas in a more northerly direction west of the launch rails. For reference, Osvaldo recorded the following coordinates for one of the alphas found: 35* 21′ 16.83″ North, 117* 48′ 50.03″ West.

Using the local wildflowers, Larry marks the location of a newly recovered alpha from the last MTA launch event

The 2019 RRS symposium was the next topic. We have over 300 Eventbrite tickets sold at the time of the meeting. The symposium has confirmed a full roster of speakers including AFRL Edwards AFB, Northrop-Grumman, USAF SMC. We decided not to hold the panel discussion this year. The symposium will start at 8:45AM on Saturday, April 27th.

Frank Miuccio goes over the preparations for the 2019 RRS symposium to be held on Saturday, April 27, 2019

The Ken Nakaoka Community Center in Gardena will allow us to set up the night before (4/26/19) at 7pm until they close at 9pm. There’s a lot of work to be done and we hope all of our membership can come out on Friday and help us with setting up tables and hanging the sign outside.

We also hope all of our membership can help at the symposium on Saturday (4/27/2019) as well. The Ken Nakaoka Community Center opens at 8AM, we will have just a little bit of time to get ready before the event begins at 8:45am with our RRS president, Osvaldo Tarditti, giving the introductory presentation.

RRS member group photo from last year’s 75th anniversary symposium (1943-2018)

The next topic of discussion at the April 2019 meeting was facility improvements at the RRS MTA. The society has decided to invest in upgrading our blockhouse and building a new restroom facility at the site for better creature comfort for the increasing number of guests we’re having each year. Osvaldo has been working up the plans for these two facility improvements and will get bids very soon.

We also hope to solicit donations from the public at the symposium to help the society reach our goals for these facility improvement projects. To anyone wishing to make a monetary donation to the RRS, you can use the “DONATE” button on the RRS.ORG homepage which connects to Paypal. Please leave us a note and accept our thanks. The society is striving to improve our facilities as we prepare to have more events this year.

Osvaldo also told us more about the RRS participating with CALFIRE in their review of the state laws governing amateur rocketry. Members of the Friends of Amateur Rocketry (FAR) organization have also been working with CALFIRE on this important committee. It is the goal of the RRS to inform the public and governing agencies on ways to make the law reasonable, practical and just to the amateur rocketry as we uphold our commitment to public safety. CALFIRE has been very supportive of our hobby and we are building stronger relationships with the State of California and our fellow rocketry organizations.

Dave Crisalli (RRS member), Larry Hoffing (RRS events coordinator), Ramiro Rodriguez (CALFIRE) and Osvaldo Tarditti (RRS president) at the RRS MTA launch event on 4/6/2019

Discussion on our last topic on the agenda was about the RRS’s participation with the base11 project. We were not able to talk about this subject in much detail as closing time had fast approached. As an educational non-profit group, the RRS has a charter to support university groups. The base11 project is very ambitious in its goal of student-run teams building and flying a liquid rocket to an altitude of 100 km or higher. This multi-year program will be a challenge on many levels both financial and technical. The RRS is happy to support the base11 Space Challenge at the RRS MTA.

The RRS is proud to support teams for the Base11 Space Challenge

The remaining agenda topics will be covered in next month’s meeting including the quarterly progress update on the SuperDosa project and the RRS partnership with Tomorrow’s Aeronautical Museum (TAM).

The Reaction Research Society meets the 2nd Friday of each month at the Ken Nakaoka Community Center in Gardena, California, at 7:30pm.

The RRS is very exciting about the projects we have planned for this summer. Our next monthly meeting will be Friday, May 10th, 2019 at 7:30pm.

Liquid Rocket Components: Pyrotechnic Valves

by Tom Mueller

Editor’s Note: This is a reprinting of the original article written by RRS member, Tom Mueller on the subject of pyrotechnic actuated valves around 1995 (?). He mentions the build of two different rockets (the XLR-50 and the Condor) and a hypergolic rocket he intended to build after this article was written. We hope to gather more photos and details about these rockets and display them in future improvements to this posting. For now, please enjoy the subject matter as the information is very relevant today to amateur builders of liquid rockets. The RRS has been very active lately in re-exploring liquid rockets. The society thought this would be a timely and interesting subject to share with our readers.

For any questions, please contact the RRS secretary, secretary@rrs.org

For an amateur rocketeer seeking to build a liquid rocket, one of the most difficult components to obtain or build are remotely operated valves. A liquid rocket will require at least one valve to start the flow of propellants to the combustion chamber. In the two small liquid rockets I have flown in the last year or so, both used a pyrotechnic fire valve located between the pressurant tank and the propellant tanks. The propellants were held in the tanks by burst disks (or equivalent) in the propellant run lines. When the fire valve was actuated, the sudden pressure rise in the propellant tanks blew the burst disks, allowing propellant to flow to the injector. This method of controlling the flow to the rocket allows the use of only one valve, and eliminates liquid valves.

In the case of the first rocket, the XLR-50 which flew in October 1993, elimination of the liquid valve was important because the oxidizer was liquid oxygen, and a small cryogenic compatible valve is very difficult to construct.

For the second rocket, which flew in October 1994, the small size prevented the use of liquid valves. In fact, the single pyro valve I used was barely able to fit in the 1.5 inch rocket diameter. In this article I will describe the design of the valves that were used on these two vehicles, and variations of them that have been used in other rocket applications.

FIGURE 1: XLR-50 pyro-technic “fire” valve

The valve shown in Figure 1 consisted of a stainless steel body with a 0.375 inch diameter piston. The O-rings were Viton (material) and the squib charge was contained in a Delrin plastic cap. The Delrin was used to prevent shorting of the nichrome wire, and also to provide a frangible fuse in case the squib charge proved to be a little too energetic. In practice, I’ve never had the Delrin cap fracture.

The inlet and outlet lines to the tanks were silver brazed to the valve body. The valve was tested many times at inlet pressures of up to 1000 psi without any problems, other than the O-rings would need replaced after several firings due to minor nicks from the ports. To help alleviate this problem, the edges of the ports were rounded to help prevent the O-ring from getting pinched as the piston translates. This was accomplished using a small strip of emery cloth that was secured in a loop in one end of a short length of 0.020-inch stainless steel wire. The other end of the wire was clamped in a pin vise which in turn was chucked in a hand drill. As the wire was rotated by the drill, the emery was pulled snugly into the port, where it deformed into the shape of the inlet, and rounded the sharp edge. I used WD-40 as a lubricant for this operation, allowing the emery to wear out until it would finally pull through the port. I repeated this process a few times for each port until the piston would slide through the bore without the O-rings snagging the ports.

Another requirement is to lubricate the O-rings with a little Krytox grease. This helps the piston move freely and greatly reduces the problem of nicked O-rings.

FIGURE 2: Fire valve for a micro-rocket

The pyro valve I used in the 25 lbf thrust micro-rocket that was launched in October of 1994 is shown in Figure 2. This valve was identical in operation to the XLR-50 valve, with the major difference being its integration into the vehicle body. The valve body was a 1.5 inch diameter aluminum bulkhead that separated the nitrogen pressurant tank and the oxidizer tank. Because of the very small diameter of the rocket, the clearances between ports and O-rings were minimized, just allowing the valve to fit. The fuel outlet port was located at the vehicle center, providing pressure to the fuel tank by the central stand pipe that passed axially down the oxidizer tank. The piston stop was a piece of heat-treated alloy steel that was attached to the valve body by a screw. This stop was originally made from aluminum, but was bent by the impact of the piston in initial tests of the valve. The black powder charge in the Delrin cap was reduced and the black powder was changed from FFFg grade to a courser FFg powder, but the problem persisted. The stop was re-made from oil hardening steel and the problem was solved. In this application, the port diameters were only 1/16 inch so only a small amount of rounding was required to prevent the O-rings from getting pinched in the ports. The valve operated with a nitrogen lock-up pressure of 1000 psi.

FIGURE 3: Fire valve for Mark Ventura’s peroxide rocket

A more challenging application of the same basic valve design was used for the fire valve of Mark Ventura’s peroxide hybrid, as shown in Figure 3. This was the first application of this valve where liquid was the fluid being controlled, rather than gas. In this case the liquid was 85% hydrogen peroxide. The second difficulty was the fact that the ports were required to be 0.20 inch in diameter in order to handle the required flow rate. The valve was somewhat simpler than the previous valves in that only a single inlet and outlet were required. The valve body was made from a piece of 1.5-inch diameter 6061 aluminum, in which a 1/2-inch piston bore was drilled. The piston was also 6061 with Viton O-rings, which are peroxide compatible. The ports were 1/4-inch NPT pipe threads tapped into the aluminum body. The excess material on the sides of the valve was milled off, so that the valve was only about 3/4 of an inch thick, and weighed only 4 ounces. Even though the piston size was 1/2 inch, the same charge volume used in the 3/8 inch valves was sufficient to actuate the piston.

In testing the valve with water at a lock-up pressure of 800 psi, I was pleased to find that even with the large ports, O-ring pinching was not a problem. One saving factor was that the larger size of the ports made it easier to round the entrances on the bore side. The valve was tested with water several times successfully before giving it to Mark for the static test of his hybrid.

The only problem that occurred during the static test of hybrid rocket was that the leads to the nichrome wire kept shorting against the valve body. Three attempts were made before the squib was finally ignited and the engine ran beautifully. I have since been able to solve this problem by soldering insulated 32-gauge copper wire to the nichrome wire leads inside the Delrin cap. In this way, I can provide long leads to the valve with reliable ignition.

My next liquid rocket is a 650 lbf design that burns LOX and propane at 500 psia. This engine uses a Condor ablative chamber obtained from a surplus yard. For this reason, I call it the Condor rocket. This rocket uses a scuba tank with 3000 psi helium for the pressurant. I decided to build a high pressure version of my valve as the helium isolation valve for this rocket. When firing this rocket, just prior to the 10 second count, this valve will be fired, pressurizing the propellant tanks to 600 psi. I assumed going in to this design that the O-rings slipping past a port simply wasn’t going to work at 3000 psi.

At these pressures, the O-ring would extrude into the port. In order to get around this problem I came up with the design shown in Figure 4.

FIGURE 4: High pressure helium valve for Condor rocket

For this valve, the O-ring groves were moved from the piston to the cylinder bore of the valve body, so the O-rings do not move relative to the ports. The piston is made from stainless steel with a smooth surface finish and generous radii on all of the corners. The clearance between the piston and the bore was kept very small to prevent extrusion of the O-rings. The valve operation is similar to the one shown in Figure 3, and the valve body is made in the same way except female AN ports were used rather than NPT ports. When the valve is fired, the piston travels from the position shown in Figure 4a to that shown in Figure 4b. During this travel, the inlet pressure on the second O-ring will cause it to “blow out” as the piston major diameter translates past the O-ring groove. The O-ring is retained around the piston, causing no obstruction or other problems. This valve has been tested at 2400 psi inlet pressure with helium and works fine. It will be tested at 3000 psi prior to the first hot fire tests of the Condor rocket next spring.

As a side note, essentially an identical valve design as the one used on the Condor and Mark’s valve is a design shown in NASA publication SP-8080, “Liquid Rocket Pressure Regulators, Relief Valves, Check Valves, Burst Disks and Explosive Valves”.

A second pyro valve is used on the Condor system as shown in Figure 5. This valve is used to vent the LOX tank in the event of a failure to open the fire valve to the engine.

FIGURE 5: Emergency vent valve for LOX tank, Condor rocket

When the propellant tanks are pressurized by the helium pyro valve, the LOX tank auto vent valve (shown in Figure 6) closes. If the engine is not fired after a reasonable amount of time, the LOX will warm up, building pressure until something gives (probably the LOX tank). The pyro valve shown in Figure 5 is used as the emergency tank vent if the engine cannot be fired. The valve body is stainless steel with a stainless tube stub welded on for connection to the LOX tank. This valve has been tested to 800 psi with helium and works fine. In this case, some ‘nicking’ of the O-rings can be tolerated because the O-rings are not required to seal after the valve is fired. The ports in the bore are still rounded, however, to prevent the O-rings from getting nicked or pinched during assembly of the valve.

Even though it is not a pyro valve, I have shown the LOX auto-vent valve in Figure 6 because this design has proven to be very useful for venting cryogenic propellant tanks without requiring a separately actuated valve or control circuit. The valve uses a Teflon slider that is kept in the vent position as shown in Figure 6a.

This allows the tank to vent to the atmosphere, keeping the propellant at its normal boiling point. When the helium system is activated, the pressurant pushes the slider closed against the vent port, sealing off the LOX tank, as shown in Figure 6b. An O-ring is used around the slider to give it a friction fit so the aspiration of the LOX tank does not “suck” the slider to the closed position. This problem happened to David Crisalli (fellow RRS member) when he scaled this design up for use on his 1000 lbf rocket system. I have used this design on the LOX tank of my XLR-50 rocket, which used a 1/4-inch diameter slider, and on the Condor LOX tank, which uses a 1/2 inch slider. In both cases the vent valve worked perfectly.

FIGURE 6: Automatic LOX tank vent valve

The main fire valve on the Condor rocket is a pair of ball valves that are chained together to a single lever so that both the fuel and oxidizer can be actuated simultaneously for smooth engine startup. For static testing of the rocket, I will use a double-acting air cylinder to actuate the valves. For flight, however, I plan to use a pin that is removed by an explosive squib to hold the valve in the closed position. When the squib is ignited, the pin is pulled by the action of the charge on a piston, allowing the valves to be pulled to the open position by a spring. This method may not be very elegant, but it is simple, light, and packages well on the vehicle. David Crisalli has successfully employed this technique on his large rocket.

That covers the extent of the pyro valves I have built or plan to build so far. In the next newsletter, I will present the design and flight of the small hypergolic propellant rocket that used the valve shown in Figure 2.