Limitations of fire safety tests and application of their results to your particular process might explain some accidents you can't figure out. So, let's consider some ways these tests can get you in trouble. For liquids and vapors, the tests include ones to determine: autoignition temperature (AIT), e.g., ASTM International's E659; flash point, e.g., ASTM D 2883; upper and lower flammability limits, e.g., ASTM E681; and sustained burn temperature (i.e., fire point), e.g., 49 CFR 173 Appendix H and ASTM D4206.
The AIT of a pure compound is affected by its concentration in a vapor cloud, containment, e.g., by steel or glass, the presence of catalysts and, especially, the ability of the observer to detect a reaction. For example, the AIT of benzene is 200°F higher in an iron flask compared to the same reaction in glass. With regard to reaction detection, some testing protocols have proven faulty — ASTM has withdrawn D286 and D2155 (for petroleum products, not the current D2155 for aircraft hydraulic fluids). Unfortunately, many older material safety data sheets (MSDSs) contain data likely based on these inaccurate tests. From my experience with combustion testing of rocket propellants, I can assure you it's difficult to detect ignition. We used high-speed cameras.
Obviously, the lower the temperature of the reaction, the lower the allowable temperature to prevent ignition of a vapor cloud. Let's assume the old test gave an AIT of 460°F. The allowable operating temperature for electrical equipment must be less than the AIT; according to the National Electric Code (NEC) Article 500, Table 500.8, we would require a temperature code of T2C, which has a maximum surface temperature of 446°F. Now, suppose the new test shows an AIT of 448°F — is this too close for comfort?
The actual AIT is influenced by factors such as temperature, total pressure, vapor pressure, viscosity and surface tension. The value listed in an MSDS is based on standard conditions — not what happens when a boiling hydrocarbon spills out of a valve stem or a furnace coil leaks.
And that AIT also is for a pure compound. Minute concentrations of other chemicals can affect the formation and ignition properties of a vapor cloud, so accurately knowing the composition is paramount to avoid risk. Testing procedures should require careful monitoring of composition. Testing may be a one-time deal. Heating too quickly might drive away the very components that are more easily ignited.
Of course, you must make some assumption about normal conditions. What about abnormal circumstances? Small wonder some accidents can't be explained.
Let's move on to the flash point. This is temperature at which a vapor cloud produced by a liquid will initiate a flame if an ignition source is present; that temperature might not sustain a flame, though. This is the difference between the AIT and the flash point: the AIT is spontaneous ignition without an ignition source while flash point requires a source of ignition. The AIT always is higher than the flash point. All of the comments on the AIT also apply to the flash point but now two new variables come into play: the ignition source and its interaction with the vapor cloud.
Let's start with the four common tests: 1) closed cup, ASTM D65, for 194°F or less; 2) Pensky Martens closed cup, ASTM D93, for liquids 194°F and above; 3) open cup, ASTM D1310; and 4) Cleveland open cup, ASTM D92. Other specialized flash point tests exist but these four dominate, with Pensky Martens most common. An open cup test produces values that typically are 9–18°F higher than a closed cup one. Regulatory agencies recognize the closed cup test as superior. Another point of contention with flash point testing is equilibrium (stirred) versus non-equilibrium tests. The liquid is well stirred in a Pensky Martens test cup. In a leak, who's to say that the lighter components won't escape first?
For additional information on flash point, see:
As for the sustained burning temperature, it's generally 18°F above the flash temperature; the burn must exceed five seconds. The same concerns as for the flash point remain and another level of complexity is added: time. Could a slow, sustained reaction start a fire at a lower temperature than measured?
DIRK WILLARD is a Chemical Processing contributing editor. He recemtly won recognition for his Field Notes column from the ASBPE. Chemical Processing is proud to have him on board. You can e-mail him at firstname.lastname@example.org