Oxygen Compatible Air

Summary

Adding compressed air directly to pure oxygen brings the danger of fire caused by ignition of the piping which becomes fuel for the oxygen. Ignition comes from hot spots inside the piping or SCUBA flask which can be caused by particle impingement and oxidative heating of accumulated fugitive oil. Sampling, analysis and specification protocols for this differ between laboratories; the differences are significant.

Background

The least expensive way to prepare oxygen enriched air (i.e. Nitrox) is to add compressed air directly to a cylinder which contains pure oxygen; this is called Partial Pressure Blending. Unfortunately, partial pressure blending is potentially dangerous. A fire in pure oxygen starts with ignition by particles hitting/scraping the piping walls as they rapidly shoot through the piping or by spontaneous combustion of an accumulated oil spot. Therefore, air to be mixed with pure oxygen must meet purity criteria which are more stringent than for normal SCUBA air.

Discovering Ignition Sources

The most accurate way to discover particulate and/or oil deposits is by rinsing the piping and analyzing the rinsate (Mil Standard 1330d). Since this is not practical in dive shops, the alternative is to check the final air that is going to be mixed with the oxygen. Discovering trace oil and particulate by air stream testing of OCA systems is not perfect, but if the flow of air is fast and long enough, this technique can discover a major filter failure and, thereby, draw attention to a probable ignition source.

Assessing the potential for an explosion by sampling and analysis of the air system requires greater attention to the design, preparation and use of the sampling kit than is true for SCUBA air sampling. If a sampling kit is not designed well for SCUBA air, it would be far worse to use it for sampling OCA. Although we can demonstrate that our kit is far superior to any other kit in the U.S. for OCA sampling, the sampling process (as well as safe oxygen dilution by OCA) it is not perfect.

Purity Specifications

A lab report that tells the client that the air is indeed Oxygen Compatible means that the lab found the air to be compliant with the OCA specification. However, there are at least 6 very different specifications in use by labs.

This, in turn, means Determining particulate and oil in Analysis of OCA is also complicated by the lack of consensus as to what specification is appropriate. The most common specifications for OCA come from AAUS, ANDI, IANTD and the US Navy. Although all four agree on lower levels of condensable hydrocarbons (0.1 mg/m3 ), a technical problem exists regarding acceptable particle size and quantities. The table below shows how these specifications differ.

The AAUS and ANDI specifications list 1 and 2 microns as the particle size not to be exceeded; however no mention ismade of quantity. Expecting an air sample taken from piping to be free of 2 microns particles is ludicrous. Think about it. Aside from the absence of experimental data to support the 2 micron theory, the specification is even far more stringent than for a Class 100 clean room. One could say "so what, at least it errs on the side of safety". But that is not only not true, but -worse - it gives one a false sense of security. Labs that say they test for 2 micron particles are - let me try to say this kindly - are fooling themselves and the dive shops that rely on that data. It is neither practical nor possible for a lab to actually do that test.

Obviously there is a safe limit of particle mass and quantity which must not be exceeded. Failing to discover experimental data supporting the AAUS and ANDI specification, we decided to rely on the results of experiments contracted by the world's best authority on handling oxygen: NASA. Our Modified OCA specification is listed below as well as here we have included our rational to allow peer review. (I have not included the criteria used by Canada or European countries; in general their acceptance criteria are significantly more stringent in some areas, and technically deficient in others.)


COMPONENT


US Navy O.C.A.


ANDI O.C.A.


AAUS
O.C.A.


IANTD O.C.A.

Our MODIFIED O.C.A.

Oxygen %

20-22

20-22

N.A.

20 -22

20-22







Fibers microns

-

-

-

-

Per NASA*

Dew Point no warmer than

- 63º F

- 50º F

- 40º F

Per client

- 64º F

Pronounced Odor

None


Our Rational for the Above Draft Version

I believe that one should keep in mind that Nitrox or Enriched Air is simply CGA Grade E air that has more oxygen, less particulate, and zero (or nearly so) oil vapor or mist. It should be clear from this that Oxygen Compatible air need only address the Particulate and Oil issues. The CO and CO2 need not be changed. Some specifications require a check of halogenated hydrocarbons because they have been frequently used to clean gauges and piping, but we have not included them in order to address the majority of OCA users.

Carbon Dioxide (1000 ppmv): Some OCA specifications above have reverted to the previous version (500 ppm) of CGA Grade E. This is probably appropriate for Deep Diving, but we have not discovered any supporting arguments for this in recreational diving. We, therefore, decided to retain the CGA Grade E criterion for carbon dioxide until other evidence is obtained.

Carbon Monoxide (10 ppmv):
This level is the same as in CGA Grade E.

Condensable Hydrocarbons (0.1 mg/cubic meter):
This seems reasonable since condensable hydrocarbons can build up over time.

Total Volatile Hydrocarbons (15 ppmv including CH4): In our experience, levels of Volatile Hydrocarbons above 10 ppmv are unusual, and in this business, unusual is generally a warning to investigate. Accepting 25 ppmv as a limit is ignoring the unusual.

Dew Point
The Dew Point of OCA is significant because of the increased tendency to rust. Rust particles have a sufficient mass and can be propelled by high flow to impact on a surface surrounded by pure oxygen. The result can be ignition and a fire. Keeping the air dry is one more step toward safe Nitrox preparation.

Particulate* (a particle count):
101 - 250 µm (93); 251 - 300 µm (3); > 300 µm (0). Although other specifications use a 1 or 2 micron limit, we have sought but have not discovered any evidence to support that idea. We make note of particles smaller than 100 µm, but we actually count and limit larger particles as listed above. The limits we show have been adopted from the NASA Safety Manual.

Fibers are defined as < 25 µm diameter particles having a >3:1 aspect ratio:
0 - 500 µm (20); 501 - 1000 µm (3); >1000 µm (0). The sizes we show are adopted from the NASA Safety Manual.

Values for particulate and fibers were taken from NASA's Oxygen Cleaning Specifications which involve analysis of the rinsate from 1 sq ft of surface area. This is basically an absolute number or total count rather than a concentration per volume of gas. We, therefore, have also considered particle limits as absolute or a total number. By sampling at a velocity higher than during oxygen blending, we hope to approach the effectiveness of rinsate sampling.

It is also important to recall that the greatest danger of an oxygen fire is at the point where the pure oxygen and the compressed air come together. There are many issues regarding the preparation of Nitrox.

Laboratory Analysis of the OCA Sample

Laboratory analysis of the OCA sample is far trickier than for SCUBA air. For example, when checking normal compressed air for Oil mist + Particulate, one simply weighs the filter before and after sampling. This allows some room for error without diver endangerment (providing that the sample was properly taken). For OCA, however, this simple test - even with more sensitive balances - is not very reliable; measuring non-volatile hydrocarbons below 0.1 mg/cubic meter introduces new considerations more significant in OCA than in SCUBA air. Although 0.1 mg/cubic meter seems like a respectably low quantity, even trace levels of sticky debris can build up to a point where they can become an ignition source if it breaks loose. This possibility makes sensitive, specific, and reliable tests even more important. Moreover, aside from hydrocarbons, determining acceptable levels of solid particulate further complicates things because ignition factors include mass and morphology.

With all these factors in mind, we employ several techniques for OCA analysis. We use microscopy for particle analysis (100 microns or greater), and extraction for GC analysis. Infrared Absorption or GC analysis are the only reliable methods for measuring condensable hydrocarbons on a filter.

Oxygen Compatible Air: Sampling Kit

With this appreciation of basic principles, we have made modifications in our sampling kit preparation and procedures to allow our standard Recreational Diving air sampling kit to be use for OCA as well as SCUBA air sampling.

  • The air must be sampled at a high velocity in order to loosen up any particulate debris which could result in a hot spot from particle impingement. The air path within the sampling kit must be very short, and polymeric tubing should not be used. (Polymer tubing attracts and, thereby, can remove oil and debris from the pathway during sampling; this prevents their discovery during sampling.
  • In order to avoid a false report, the analytical filter and the sampling equipment must be tested for particles and fibers before it is sent out.