It’s no secret that flash point tests are some of the most common tests performed by the Flammability team at Fauske and Associates. After all, these tend to be some of the most cost-effective ways to evaluate the flammability of a material. The flash point test’s wide range of applicability combined with its low barrier of entry has led to its adoption as a high-priority safety measure by many regulatory and standardization organizations, such as the National Fire Protection Association (NFPA) and the Occupational Safety and Health Administration (OSHA).
The basic steps in a flash point test are similar across all flash point methods. A small amount of sample is inserted into a cylindrical test cup and is subsequently heated. When it reaches the desired testing temperature, a small pilot flame is introduced to the vapor space of the sample cup. If the flame propagates instantaneously across the surface of the sample, this is referred to as a flash. The lowest temperature at which a flash is observed is referred to as the sample’s flash point. Knowing the flash point of a material is very useful because it indicates the lowest temperature at which a liquid can produce enough vapor to form an ignitable mixture with air, essentially determining its flammability and allowing for proper safety measures to be taken regarding storage, handling, and transportation to prevent fires and explosions.
Many methods have been devised to produce the best results for the widest variety of samples. These methods often change several variables, which include (but are not limited to) the heating method, sample quantity, and stirring speed. Some of the more impactful variations are well represented by the three most common closed cup flash point methods used by Fauske and Associates. These three methods and their appropriate testers are:
- Tag Closed Cup Tester via ASTM D56
- Pensky-Martens Closed Cup Tester via ASTM D93
- Small Scale Closed Cup Tester via ASTM D3828
Study Overview
While there are many variations between the testers, some of the most notable are listed in Table 1. This article does not aim to show the direct effect of any one variable on the flash point; instead, it focuses on the specific combinations of variables found in each tester and their appropriate ASTM test standard(s).
These three testers all have well-defined use cases for samples with atypical properties, although they are often thought to be comparable when it comes to more cooperative samples. Here we put that hypothesis to the test.
This article focuses on testing three relatively cooperative materials using the three testers listed in Table 1. Specifically, the three samples tested are pure N,N-dimethylformamide (DMF), 10% Ethanol in deionized water, and 8% Isopropanol in deionized water. The listed dilutions are in volume percent. All these samples are common verification materials, are readily available, and flash in the applicable temperature range of each tester used, as listed in Table 1.
Initial Results
In accordance with the appropriate ASTM methods, all resulting flash point temperatures in Table 2 were adjusted for the atmospheric pressure and rounded to the nearest 1°C.
Initial Observations & Analysis
To preface the findings, note that the hypothesis being challenged through this study is that each flash point tester should return nearly the same value when testing the same sample.
The results for DMF do not disprove this hypothesis. The difference between the reported flash points from the three methods is relatively minor, with the highest reported value coming in at only 3°C above the lowest. Observations reported by the operator were similarly consistent between methods; the sample showed a large orange halo around the pilot flame as it increased in temperature and ignited quickly upon reaching its flash point temperature.
The method giving the highest reported flash point for DMF was the Pensky-Martens method. This is an unsurprising result, given the faster temperature ramp rate of the tester as shown in Table 1 such that the sample often has less time to thermally equilibrate. While the stirring inherent to the tester lessens the effect of this quick heating, the DMF sample still shows the 3°C difference. This result is in-line with our experience.
The results of the other two samples, however, showed larger differences. In both cases, the results on the small-scale tester were much higher than that of both other testers, with the 10% Ethanol flash point having a notable 7°C difference between the Small Scale and Tag results. These results, which were consistent in retesting of the samples, challenge the hypothesis.
The two “troublesome” samples, 8% Isopropanol and 10% Ethanol, may be impacted by their large percentages of water. When water’s high heat capacity is inherent to the sample, the one-minute isothermal hold mandated by the ASTM standard may be insufficient to allow the sample temperature to adequately equilibrate prior to introducing the pilot flame into the vapor space. This insufficient temperature stabilization may inhibit the evolution of flammable vapors into the headspace of the test cup, effectively delaying the flash point until a higher temperature can overcome this increased heat capacity.
Fortunately, there is a precedent for testing troublesome samples in the Small Scale tester. In ASTM E502, Standard Test Method for Selection and Use of ASTM Standards for the Determination of Flash Point of Chemicals by Closed Cup Methods, the standard references a modification in the test procedure to allow six minutes for the sample temperature to equilibrate instead of the usual one-minute isothermal hold. The written intent is to use this modification for samples with high viscosity as well as samples that flash while solid; thus, in this instance, it is a minor deviation from the standard.
We decided that the hold time modification should be performed on all samples in this study. The results of this, as well as the previously included one-minute hold results from Table 2, are shown below in Table 3.
Additional Trials at a Six-Minute Hold
The results shown in Table 3 speak to the effectiveness of this increased hold time on the flash point of the diluted samples. While the pure DMF had no notable change, allowing the other aqueous samples to better equilibrate prior to the introduction of the pilot flame ignition source had a measurable impact on the reported flash point values. In fact, for both 10% Ethanol and 8% Isopropanol, the six-minute hold time cut the difference between their small scale and Tag results in half, down to a 3°C difference for 8% Isopropanol and a 4°C for 10% Ethanol.
When observed together, the results listed in Tables 2 and 3 show only small differences in the effective results of each tester. DMF and 8% Isopropanol showed a maximum difference of 3°C and 10% Ethanol showed a maximum difference of 4°C, ensuring that any method listed here can achieve a similar result on these materials.
Why Choose One Method Over Another?
Given the results of this study, it becomes clear that some samples provide more consistent results when tested according to different methods. With that being said, how does Fauske & Associates choose which method to perform when the customer defers the choice to us?
The answer lies in the strengths of each tester. As listed in Table 1, there are a variety of different variables to consider, and each method finds special cases where its characteristics exceed the other testers.
The Tag method’s ramp rate, for example, is intentionally much slower than that of other methods. This slow ramp rate, combined with the ease of use and access to internal cooling, makes it optimal for samples that are expected to flash close to or below ambient temperature. It also features the highest precision among these methods, as it gives the sample plenty of time and space to generate vapors. Its major shortcoming, however, is its limited applicable temperature range. As shown in Figure 1, the Tag method has a relatively narrow testing range of -37 to 93°C, severely limiting the variety of samples it can test.
The Pensky-Martens method lacks the sub-ambient capability and gradual temperature rise rate of the Tag method but makes up for it through its much broader temperature range. Spanning from 40°C to 370°C, the Pensky-Martens tester lends itself to a wide range of samples and testing conditions. Its included stirring also makes it valuable for samples with high viscosities and/or components that settle out. These traits also make it useful for performing screening tests, in which a sample is quickly run through the Pensky-Martens test method to confirm its traits.
The Small Scale method has a wide range of useful qualities. Its small sample size requirement lends itself well to inherently safer design principles by minimizing the amount of hazardous sample needed for testing. Its stainless-steel test cup, in contrast to the other two methods listed here, is also resistant to corrosive chemicals that a brass test cup is vulnerable to. Moreover, it serves as the only one of these methods that can test solid samples and is the most consistent for highly viscous samples.
Conclusion
The initial trials showed a clear inconsistency across the three tested methods. While the Pensky-Martens and Tag testers displayed relatively similar flash points for all three samples, the two diluted samples experienced notably higher flash points on the Small Scale tester. Extending the isothermal temperature hold time for the Small Scale tests to 6 minutes lessened the effect of the increased water content, bringing every tester back to results that are within a few degrees of each other.
With this modification in place, there is only a minor difference in results between testers for samples that are similar to those tested here. The choice of which tester to use on a material is then dependent on the properties of the material and the associated traits of each tester.
There are many more variables to examine in follow-up studies. A few notable variations could be using other ignition sources, such as an electrical discharge in a glass flask, or halogenated samples, which often display atypical properties when tested using these flash point methods. Fauske & Associates hopes to continue presenting this research in the future.
For more information on flammability testing services contact flammability@fauske.com.