Chris Schultz

Abstract

Flash points are a measurement of the flammability of vapors. [1] Flash points are a necessary safety standard to help control and recommend the safest way of transport and use of chemicals for manufacturers, and industries. There are numerous standards for finding flash points and different agencies that standardize flash points, from U.S. to European, and International agencies. Flash points were originally introduced in 1862 [2] to control quality and safety of kerosene products, and indeed are still mainly used for the safety of manufacturing, transport, and use of petroleum products and fuels. There are two main types of apparatus for testing experimentally the flash point of materials, closed and open cup testing. In more recent years there have been numerous studies into finding flash points without extensive experimental testing, using boiling points, viscosity, and Le Chatelier’s Principle. Unfortunately experimental results can vary and are not reliably repeatable, and while numerous automated devices have been developed, scientists are still attempting to find more repeatable and reliable ways to determine flash points of liquids. The aim of this paper is to show the importance, backgrounds, and continues research into flash points.

Flash Points and Vapor Pressures

The flash point of a liquid is the temperature at which the vapors above a liquid create a mixture with air sufficient to be ignited by a flame from an ignition source, and are standardized to 101.3 kPa. Flash points are not to be confused with fire points/lower flammability limits, which are the point at which vapors will continue to propagate a flame even after an ignition source has been removed. Flash points are also often confused with auto ignition points, the point at which a vapor will ignite without an ignition source. [3] Vapor Pressure is the measure of the amount of vapor fumes of a liquid in an enclosed system. Vapor pressure is an important concept to understand when discussing flash points; flash points measure when a vapor has interspersed throughout an air mixture enough to be lit by the application of an ignition source. Vapor pressures are the pressure induced upon a container from vapors produced from the evaporation of liquids or the sublimation of solids. As temperatures increase vapor pressures also increase as more atoms or molecules have enough energy to break intermolecular forces binding them to the rest of their liquid or solid structure. Generally higher vapor pressures indicate lower flash points of molecules and mixtures, as higher vapor pressures allow gasses to permeate the surrounding air.

General Methods for Finding Flash Points


Methods for testing flash points are broken down into two main methods; open cup and closed cup. In the open cup method, liquids are heated in a an open cup, and an ignition source is periodically passed over the cup, exposing vapors to ignition. The flash point is determined to be the temperature at which the vapors ignite. Open cup apparatus and testing is designed to show how liquids will act if spilled as generally happens in in transportation accidents. Open cup methods generally show higher flash points, and have less repeatable data. This is due mainly to the fact that vapors have a tendency to disperse more than closed cup methods, and that any small drafts in the area where the experiment is performed can make the results less accurate. [4] Closed cup methods involve a closed container which is heated, and the lid removed after nearing desired temperature and an ignition source applied. Closed cup methods more accurately depict the way solutions will react in closed containers, and so are used more often for the transport and storage of chemicals. Closed cup methods are generally more reliable, and normally produce lower results for flash points, and thus are considered the safer measure of a chemical’s flash point due to the lower values found. One final reason for the preferential use of closed cup methods is the ability of samples to reach vapor pressure equilibrium with the surrounding air before being exposed to ignition. Both general methods are susceptible to incorrectly read tests and non-homogeneous samples especially from undertrained users. Below is depiction of the two methods showing the general difference between the two methods.



open_close_cup.png


Testing Methodology and Setaflash as a General Standard

Some of the most common tests and their applicable ranges for flash points are the [5] Pensky-Martens Closed Cup 60C-190C,[6] Cleveland Open Cup 79C-400C Tag Open-Cup 93C and above (no upper bound specified), ASTM D3828 Small Scale Closed Cup Tester Setaflash -30C-300C. The Setaflash tester clearly has the largest range of flash points that it is able to classify from samples. Below is a table showing the flammability and combustibility classifications of liquids which is part of the main importance and reasoning behind determining flashpoints. [7]


Flammability classification of liquids


DOTClus DOLCIBBS Flash point range
Flammable <100°F(<37.8°C)


Flammable FlammablelA <73°F(<23°C) (boilingpoint<lOO"F)

Flammable FlammableIB <73"F(boilingpoint <.100°F)

Flammable FlammableIC >73"Fand<100°F Combustible

Combustible >.l00°Fand<200°F(93.3°C)

CombustibleII >.100°Fand<140°F(60°C)

Combustible CombustibleIlIA >.140°Fand<200°F

There are several different agencies that standardize flash points. The forefront runner of these is Joint ISO/CEN Working Group on Flash Point. Below is a table from Mike Sheratt, Chair of Joint ISO/CEN Working Group on Flash Point [8]

chart_3.png

“The Small Scale (Setaflash) Closed Cup test is specifically identified by the following test methods: ASTM D3278, ASTM D3828, IP303, IP523, IP524, EPA 1020 A and B, ISO 3679 and ISO 3680.” [8]
As the chart clearly depicts the Small Scale testing method is the most widely accepted method for determining flash points of chemicals. [9] Setaflash Small Scale testing method and apparatuses are used throughout almost all testing methods for their highly reproducible results in comparison to other flash point testing methods. They also have relatively short testing times as low as one or two minutes in comparison to tests that can run up to two hours. Setaflash has also had over thirty years of trust from the scientific community, and is generally accepted to be the most accurate method of finding flash points.
Flash points as a factor for determining explosive and fire dangers from chemicals is extremely important, as the Department of Health, Safety, and Environmental Engineering states “In the chemical industries the three most commonaccidents are due to toxic releases, fires, and explosions,with most occurring in the form of fires andexplosions.” [10] Unfortunately, there are some liquids that seem to defy classification through the standard classifications systems with flash points. [11] Liquids such as consumable alcohols are often not checked for flash point, and even though flammable are not classified as flammable due to low incidence of accidents. More interestingly Melvin Gerstein and William B. Stine researched basic inhibitors and found that they could make a liquid display no flash point. [10]Inert gasses/liquids are used to lower concentrations of flammable gasses to below the lower explosive limit, though inert gasses/liquids often create violent explosions at very high temperatures. This makes certain mixtures even more dangerous because although the liquids would show no flash point, they were still flammable, and hence more dangerous as they could be classified incorrectly as nonflammable or explosive. Explosions obtained from these mixtures are often less predictable and more violent as well.
One particular research study into the safety of miners found that flash points did not always accurately predict the flash point of liquids. [12] In addition the state of the substance is important as sprays and foams were a problem, foams are easily lit at lower temperatures than the normal flash point of their purely liquid form. This lead to safety concerns of flammable liquids when shaken or overly disturbed as highly flammable foams could form at the top layers of the explosive liquid, producing a deceptively low lower flammable limit, and not acting as predicted by flash points. Sprays were also found to affect the actual usefulness of flash points, as sprays with coarser droplets as opposed to finer particles found in gasses had a tendency to fall towards flame, and thus propagate flame more easily than would be assumed by flash point testing.

General Background in Liquid Flammability

One of the important concepts to understand in the study of flammability of liquids is the science behind the ignition of gasses and the danger posed by this. To pose a significant danger gasses must not simply ignite, but also continue to propagate the flame throughout the gas. [13] This requires the flame from the gasses to be hot enough to cause the subsequent layer of surrounding gas to burst into flames and continue the propagation of flame. An interesting phenomenon that can occur when gasses are closely confined and propagation is incredibly fast is that flames do not complete combust, leading to incompletely oxidized by-products and a phenomenon known as cool flames. Cool flames is a term used when the vapors ignite, but each layer or “shell” [13] of gas does not complete combustion, and hence do not reach the uppermost temperature for their flames, or the standard explosive results. [13] This shows a midway between LEL (Lower ExplosiveLimit) the limit at which gasses will ignite and propagate, very closely related to flash point, and UEL (Upper Explosive Limit) at which point levels of vapors actually prohibit the propagation of flame by suffocating the flame through lack of oxygen. Both limits and flash point tend to fluctuate sometimes greatly as pressure and temperature are both increased. [10] Some gasses commonly used as inhibitors are argon, carbon dioxide, and nitrogen.

Le Chetelier’s Principle

Almost all mathematical ways for finding the flash points of chemicals use Le’ Chatelier’s Principle, and so a brief explanation seems necessary. [14] Le Chetelier’s Principle is the fact that when products and reactance are left together in a reversible reaction equilibrium will be established, and if it is disturbed will gradually return to equilibrium. This simply means that if the number of products is increased in equilibrium by adding more products, then some of these new products will return to reagents, and vice versa. Eventually the two will reach the same equilibrium where the forward and back reactions occur at the same rate, and ratio of one to another, as previous to adding the additional amount of product to the mixture.

Mathematically Determining Flash Point

Many scientists have studied ways to predict Flash points, using the makeup of molecules, viscosity of chemicals, and the boiling points of liquids. [13]In 1982 a method for mathematically finding flash points using UNIFAC (Universal Functional Activity Coefficient) was developed by. Their argument shows initially the flash point of a pure liquid can be written as “P,"/Li = 1” [15] in other words the vapor pressure of a purely ignitable liquid divided by the lower flammability limit when equaling one shows the flash point of the liquid. They go on to use a principle from Le Chatelier stating that the summation of partial pressures combustible components in a compound will replace the vapor pressure of the compound as a whole to find flash point. “∑Pi / L i= 1” [15]This is relying on the fact that atmospheric pressure has little effect at standard pressure 101.3 kPA. When flash points remain lower, and temperature change is low an equation from Zabetakis (a fire safety engineer who produced data for flammability of liquids in 1965) may be used to correct these small changes. The authors go on to show that at lower pressures the mixture of air and the ignitable source may be thought of to act like ideal gasses. Unfortunately many industrial uses of chemicals is at high temperature and pressures, and as temperature and pressure increase these results land farther and farther away from experimentally determined values.
Another mathematical expression attempting to relate the flash point of chemicals to their boiling points was also developed in 1982, [16] this one by Fujii and Herman. Fujii and Herman developed mathematical formulas that helped correlate boiling point to the flash points of chemical compounds of several different types of organic compounds. Alkanes, Aromatics, Alkenes, Ethers, Ketones, Aldehydes, Acetates, Esters, Alchohols, Phenols, and Amines were all studied. [C13] The findings were as to be expected with flash points rising as boiling points rose, with one interesting find regarding double bonds. Double bonds raised boiling points but lowered flash points. This is thought to be because of the susceptibility of double bonds to the attack of free radicals, hence the lowering of flash points. Unfortunately while this study was another step forward in attempting to find a method for accurately predicting flash points, only 31 chemicals were tested, only 55% of findings where within 5°C of experimentally determined values, though 89% were within 10°C. Acceptable range from experimentally determined value as determined by this experiment was stated to be 3.7°C. This paper does not manage to develop a way to estimate the flash point of mixtures of chemicals, and also is not reliable enough to be safely used as a primary method of determining flash points, but could be used as a secondary source.
[17] Published in 2002 “A mathematical model for predicting the flash point of binary Solutions” By Liaw while not the most comprehensive of mathematical solutions for finding flash points is easily one of the most repeatable and accurate methods. Below is a list of the constants needed to find activity coefficients for binary systems. Below the constants is a table showing some of the methods that allow one to derive the activity coefficients for binary systems. [17]
Nomenclature,_chart_1_0f_2.png
nomenclature_chart_2_of_2.png
[17]

Using the equations listed above and Le Chetelier’s principle Liaw show’s that in binary systems flash points can be estimated. This is important as this is the first mathematical path for finding flash points that can be used for non-ideal gasses reliably, and so becomes much more valuable to industries.

Non

The main problem with many of the mathematical predictions of flash points is the fact that gasses do not act the same at all temperatures. Normally at low temperature and pressure gasses act ideally, with their particles moving freely past each other, and gas particles exerting forces on each other small enough to be negligible. Industrial storage and reactions often take place at high extremes of temperature, and high pressure. High temperature actually causes gasses to act more ideally, as forces between molecules are less relevant the faster atoms or molecules are traveling. Problems arise at low temperatures, and extreme pressure, as both lead gasses to act as non-ideal gasses. At high pressures the fact that molecules of gas actually do take up space causes gasses to act as non-ideal gasses as they are NOT actually infinitely compressible at high enough pressures. At extremely low temperatures the intermolecular forces between molecules become a factor, and can cause gasses to act non-ideally. When gasses begin to act non-ideally flash points fluctuate, and can change dramatically. This has helped contribute to the difficulty in producing mathematical equations that accurately predict flash points

Automated Flash Point Testing, and Applications

One of the main purposes for trying to find a trustworthy mathematical equation to find flash points from other know quantities it to aid in the automation of flash point testing.[18] In 1993 a patent was passed for one of the first automated flash point testing machines available. This machine would be placed adjacent to a pipe, and periodically sample sections of the contents of the pipe. The sample would be mixed with air, and quickly heated up in a testing chamber to determine the flash point and LFL. This would then be recorded by a microprocessor in the machine. The great benefits of having automated machinery testing flash point were the time and reliability factors. Being a machine human error was removed, and there was no time lag between removing a sample from a pipe, and sending it to a lab to be tested for flammability. Unfortunately many early models of automated flash point testing were apt to become dirty through buildup of burnt residue, and required overly frequent cleaning. An automated apparatus is still greatly useful for refineries and companies attempting to produce products that may have narrow lines of safe flash points, such as many fuels.
Research continues on trying to find more precise mathematical calculation and automated testing for various aspects of crude oil and other petroleum based products. [18] Some more recent work that has managed to correlate many different aspects of crude oil throughout different stages of refinement through software based sensors. These sensors make use of a system referred to as ANN, or artificial Neural Networks to better find qualities such as flash points of products. [20] Additionally GA or genetic algorithms are used. These genetic algorithms are run by computers normally for optimization purposes, and run off previous results to “evolve” to find a better solution to problems.
Conclusion
As flash point are such an important factor in determining flammability and explosiveness and hence safety of mixtures continued research in the field, especially in predictive models is needed. Current mathematical predictions of flash point while continuing to improve still have trouble predicting with high accuracy flash points of mixtures with combinations of inert gasses and high numbers of functional components, especially at high pressures and temperatures.









References
[1]Hanley, Brian. "A model for the calculation and the verification of closed cup flash points for multicomponent mixtures." Process Safety Progress 17.2 (1998): 86-97. Wiley Online Database. Web. 1 Dec. 2011.<http://onlinelibrary.wiley.com/doi/10.1002/prs.680170204/abstract>.
[2]Damn, Manijot, and Deoki N Saraf. "Hazard rating system for flammable and combustible liquids." Process of Safety Progress 15.4 (1996): 230-236. Wiley Online Library. Web. 1 Dec. 2011.<http://onlinelibrary.wiley.com/doi/10.1002/prs.680150410/abstract>.
[3]"Vapor Pressure." Vapor Pressure. N.p., n.d. Web. 1 Dec. 2011.<http://www.chem.purdue.edu/gchelp/liquids/vpress.html>.
[D4]Thorne, P. F. "Flash Points of Mixtures of Flammable and Nonflammable Liquids." Fire and Materials1.4 (1976): 134-140. Wiley Online Library. Web. 1 Dec. 2011.<http://onlinelibrary.wiley.com/doi/10.1002/fam.810010403/abstract>.
[F5]International Organization for Standards. "Determination of flash point -- Pensky-Martens closed cup method." International Organization for Standards. International Organization for Standards, 2011. Web. 1 Dec. 2011. <http://www.iso.org/iso/catalogue_detail.htm?csnumber=36614>.
[N6]Subcomittee. "ASTM D92 - 05a(2010) Standard Test Method for Flash and Fire Points by Cleveland Open Cup Tester." ASTM International. ASTM International, n.d. Web. 1 Dec. 2011. <http://www.astm.org/Standards/D92.htm>.
[7]Lance, R. C., A. J. Barnard, and J. E. Hooyman. "Measurement of flash points: Apparatus, methodology, applications. "Journal of Hazardous Materials 3.1(1979):107-119.Science Direct Web.1Dec.2011.<http://www.sciencedirect.com/science/article/pii/0304389479850086>.
[8] Sherratt, Mike. "Flash Point - The Definitive Test Method." All About Chemistry. Hadian Taufiq, 28 Nov. 2009. Web. 1 Dec.2011.<http://chemistry02.blogspot.com/2009_02_04_archive.html>.
[9]Sherratt, Mike. "Flash point testing - The definitive test method." ChemEurope. ChemEurope, n.d. Web. 1 Dec. 2011. <http://www.chemeurope.com/en/whitepapers/45235/flash-point-testing-the-definitive-test-method.html>.
[10] Wu, Siou-Yuan, Nung-Kai Lin, and Chi-Min Shu. "Effects of flammability characteristics of methane with three inert gases." Process of ASafety Progress 29.4 (2010): 349-352. Wiley Online Library. Web. 2 Dec. 2011. <http://onlinelibrary.wiley.com/doi/10.1002/prs.10411/abstract>.
[11]Gerstein, Melvin, and William B. Stine. "Anomalies in Flash Points of Liquid Mixtures." Industrial & Engineering Chemistry Product Research and Development 12.3 (1973): 253-255. Industrial & Engineering Chemistry Product Research and Development. Web. 2 Dec. 2011.
[12]Zabetakis, M G. "Flammability characteristics of combustible gases and vapors." Bulletin of Mines 627 (1965): 1-129. Mendeley. Web. 2 Dec. 2011.<http://www.mendeley.com/research/flammability-characteristics-of-combustible-gases-and-vapors/>.
[13]Vidal, M, et al. "A review of estimation methods for flash points and flammability limits." Process Saftety Progress 23.1 (2004): 47-55. WileyOnline Library. Web. 1 Dec. 2011. <http://onlinelibrary.wiley.com/doi/10.1002/prs.10004/full>.
[14]Clark, Jim. "LE CHATELIER'S PRINCIPLE." Chemguide. N.p., Nov. 2011. Web. 1 Dec.2011. <http://www.chemguide.co.uk/physical/equilibria/lechatelier.html>.
[15]Gmehling, Juergen, and Peter Rasmussen. "Flash points of flammable liquid mixtures using UNIFAC." Industrial and Engineering Chemistry Fundamentals 21.2 (1982): 186-188. ACS Publications. Web. 1 Dec. 2011. <http://pubs.acs.org/doi/abs/10.1021/i100006a016>.
(P16) Fujii, Atsushi, and Edward R. Herman. "Correlation between flash points and vapor pressures of organic compound." Journal of Safety Research 13.4 (1982): n. pag. Science Direct. Web. 1 Dec. 2011. <http://www.sciencedirect.com/science/article/pii/0022437582900329>.
[17]Liaw, Horng-Jang, et al. "A mathematical model for predicting the flash point of binary solutions." Journal of Loss Prevention in the Process Industries15.6 (2002): 429-438. Science Direct. Web. 1 Dec. 2011. <http://www.sciencedirect.com/science/article/pii/S0950423002000682>.
[18]White, and Et Al. "Means and method of an on-line determination of the flash point of lube oil fractions." Google Patents. Google, 5 June 1972. Web. 1 Dec. 2011. <http://www.google.com/patents?id=rOF8AAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false>.
[19] Dam, Manojit, and Deoki N. Saraf. "Design of neural networks using genetic algorithm for on-line property estimation of crude fractionator products." Computers and Chemical Engineering 30.4 (2006): 722-729. Science Direct Web. 2 Dec. 2011. <http://www.sciencedirect.com/science/article/pii/S0098135405003030>.
[20]"Genetic Algorithms." Wikipedia. Wikipedia, n.d. Web. 2 Dec. 2011.http://en.wikipedia.org/wiki/Genetic_algorithms.