Reference Jet Fuels for Combustion Testing

Transcription

1 Reference Jet Fuels for Combustion Testing Tim Edwards 1 Air Force Research Laboratory, Dayton, OH, This paper provides a summary of the composition and properties of reference jet fuels used in the National Jet Fuel Combustion Program. Additional data is provided for common alternative jet fuels. Two classes of fuels are discussed: (1) Category A fuels which represent the range of properties seen in current petroleum-derived jet fuels, and (2) Category C test fuels which have certain properties well outside of experience. Combustion test and modeling results for a number of these fuels are becoming available in the literature, with this paper serving as a detailed reference on the properties of these reference fuels. Nomenclature AFPET - Air Force Petroleum Agency AFRL - Air Force Research Laboratory ASTM - American Society for Testing and Materials (ASTM International) ATJ - Alcohol-to-jet (alternative jet fuel) C p = Heat capacity (of fuel) C x = Hydrocarbon fuel component with x atoms of carbon per molecule CRATCAF Combustion Rules and Tools for the Characterization of Alternative Fuels (program) CRC - Coordinating Research Council CTL = Coal-To-Liquid (Fischer-Tropsch) DLA - Defense Logistics Agency F-T = Fischer-Tropsch fuel processing FAA - Federal Aviation Administration (United States) FAME - Fatty Acid Methyl Ester (biodiesel) GTL = (natural) Gas-To-Liquid (Fischer-Tropsch) H content = Hydrogen content in the fuel by mass HEFA - Hydroprocessed Esters and Fatty Acids (alternative jet fuel) HOC = Heat of Combustion (MJ/kg) (net) HRJ = Hydrotreated Renewable Jet (fuel) predecessor name for HEFA IPK = Iso-Paraffinic Kerosene (Sasol) MURI = Multidisciplinary University Research Initiative MW = Average Molecular Weight of a complex fuel NJFCP - National Jet Fuel Combustion Program OEM - Original Equipment Manufacturer POSF - fuel designation, not an acronym PQIS - Petroleum Quality Information System (DLA database) P = Pressure (bar or atm) SPK - Synthetic Paraffinic Kerosene (alternative jet fuel) SwRI - Southwest Research Institute, San Antonio, Texas T = Temperature (K) UDRI - University of Dayton Research Institute WPAFB - Wright-Patterson Air Force Base 1 Principal Chemical Engineer, Aerospace Systems Directorate, AFRL/RQ, WPAFB, OH 45433; AIAA Associate Fellow 1

2 I. Introduction The ongoing effort to evaluate the performance of alternative aviation fuels has renewed the interest in the effect of jet fuel composition changes on gas turbine operability and performance. Aviation s current focus is on drop-in fuels, which are composed solely of hydrocarbons, but produced from alternative sources/feedstocks such as biomass. Aviation is not considering oxygenated jet fuel components such as alcohols or fatty acid methyl esters (FAMEs) due to their negative impacts on performance and handling. A previous paper [1] described an ongoing program to streamline the combustion evaluation of alternative aviation fuels under the umbrella of the National Jet Fuel Combustion Program (NJFCP) this paper is a significant expansion of the fuels section of that paper. Based on requirements developed in an earlier program (Combustion Rules and Tools or CRATCAF [2]), the NJFCP program has developed/acquired a suite of conventional jet fuels and test fuels to characterize the fuel sensitivity/response of combustion devices. These fuels were developed to span the range of jet fuel composition and properties that could be encountered with conventional and alternative jet fuels. The fuels are being acquired and distributed by the Air Force Research Laboratory s Fuels Branch in Dayton Ohio. AFRL has also distributed typical and average fuels to the research community in the past, and some of these fuels will be included in this paper. This paper will also include the properties of common alternative jet fuels distributed by AFRL. The test fuels and alternative fuels do not meet the all the jet fuel specifications (such as density); the test fuels were designed to explore a particular aspect of the fuel property/composition space such as boiling range or viscosity, while many of the alternative fuels have been approved in fuel specifications such as ASTM D7566 for use as 50% blends (or less) with conventional fuels. The Rules and Tools program had defined several categories of fuels to be used to characterize fuel effects on combustion, some of which are being carried forward by NJFCP: Category A - conventional fuels derived from petroleum, encompassing the range of properties typically encountered (viscosity, flash point, aromatic content, etc.) Category B alternative jet fuels found to have unacceptable combustion properties (not used in NJFCP) Category C test fuels designed to explore the edges of the jet fuel composition-property space, such as fuels being at the viscosity limit of the specification or fuels whose composition is outside of typical experience (such as cycloparaffin content) This paper includes data for Category A and Category C fuels as utilized by the NJFCP, as well as data for common alternative fuels. The data presented includes: Specification properties important to combustion: flash point, viscosity, aromatic content, hydrogen content, ASTM D86 distillation, smoke point, measured heat of combustion, measured cetane number Composition: GCxGC distribution of major hydrocarbon classes across carbon number, average molecular weight Fit-for-purpose properties - density vs T, viscosity vs T, Cp vs T, surface tension vs T, vapor pressure vs T ASTM Calculation methods for properties such as heat of combustion (ASTM D3338), hydrogen content (ASTM D3343), and cetane index (ASTM D4737) are typically not validated for alternative fuels or test fuels and thus are not included, with this paper focusing on the actual measurements. Cetane index (calculation) has been found to be very inaccurate for some alternative fuels. Note that oxygenate-free jet fuels allow hydrogen content to be converted to H/C ratio directly by making the valid assumption that the non-hydrogen portion of the fuel is entirely carbon. Also included as part of the data is the AFRL identification number, which identifies specific batches of fuel. This ID number takes the form of POSF XXXXX, where POSF is the Fuels Branch s organizational designation back in 1981 when the numbering began (with 001), thus it is NOT an acronym. The number are assigned roughly in order received, with POSF assigned in late When available, comparison data is presented from common sources, such as the Coordinating Research Council s (CRC) Handbook of Aviation Fuel Properties [3], the CRC World Fuel Sampling Program [4] (often termed the World Survey), and the Defense Logistics Agency s (DLA) Petroleum Quality Information Service (PQIS) database [5]. There is a current FAA program collating airport property data that should generate a very useful set of data available in There are a number of petroleum industry references for calculating most physical properties of petroleum fractions such as jet fuel, typically using commonly measured properties such as specific gravity/density and mean boiling point from ASTM D86 [6,7,8]. Thus, while the ASTM D86 distillation does not represent a true boiling curve for jet fuels, its use since the 1930s allows it to be correlated to a large amount of historical data. NIST has developed an advanced distillation method, and has recently published data on Category A fuels [30]. NIST has also very 2

3 recently published densities of the Category A fuels as a function of pressure up to 45 MPa [31]. The densities presented in this paper are the conventional jet fuel densities by ASTM D4052 at 1 atm. Other properties are also measured at atmospheric pressure a potential shortcoming for applications at high pressure such as aviation diesel engines with high-pressure common rail fuel injectors. The NJFCP program is also a successor to earlier DoD and NASA programs in the late 1970s (e.g., [9]), which looked at the effect of fuel composition and property changes on 1970s-vintage combustors. These earlier programs also focused on broadened jet fuel specifications that could be used to increase supply, such as increasing the jet fuel aromatic limit above 25 vol%. In contrast, NJFCP is looking at the effects on current and future combustor operability of various jet fuel composition changes that might be driven by modern alternative (bio) fuels. II. Conventional/Reference Fuels Category A A. Overview The Category A jet fuels were defined by selecting important combustion-related properties and attempting to find production fuels that would represent the range of properties of jet fuels in use today. The CRATCAF program included a detailed literature review of prior work, from which the OEMs selected three combustion-related properties that represented the variations seen in practice - flash point, viscosity, and aromatics content - that would be expected to have the greatest impact on combustor behavior. Three fuels were sought: a fuel with low flash/viscosity/aromatics ( A-1 ), average/nominal properties ( A-2 ), and high flash/viscosity/aromatics ( A-3 ). Table 1 shows the OEMselected desired properties. Cliff Moses, retired from Southwest Research Institute, referenced the Petroleum Quality Information System (PQIS) database from DLA to identify originating sources for these three fuels. It was desired that 6,000 gallons (23,000 L) be obtained for the A-1 and A-3 fuels, and 22,000 gallons (83,000 L) for the nominal A-2 fuel. After some time (and effort), suitable fuels were identified and obtained directly from refineries (many refineries have no ability to load trucks). The A-2 fuel (POSF 10325) was a Jet A procured from the Shell Mobile* refinery in June The A-1 fuel was a JP-8 fuel procured from NuStar Refining* in April The A-3 fuel was a JP-5 fuel from Valero* procured in May As shown in Table 1, the average/nominal fuel goals were met with A-2. The A-1 fuel goals were nearly met. The A-3 (JP-5) had a flash point lower than desired. (It was demonstrated at AFRL that, if necessary, the flash point of this fuel could be raised to 70 C by distilling off the lower boiling components.) Also, the aromatic goal was not met but the hydrogen content of this fuel is 13.4 mass% (due to high cycloparaffin content), which is the lower limit of JP-8 specifications and is at the low end of the jet fuel experience base. Previous programs (such as described in [9]) have shown that fuel soot production is controlled by overall fuel hydrogen content or H/C ratio for hydrocarbon fuels, rather than by aromatic level. This has been verified in engine testing where aromatic-free but decalin-rich fuels burn very similarly to fuels with 25% aromatics and the same hydrogen content. *Any identifications of commercial products within this paper is for information only and does not indicate recommendation or endorsement by FAA, AFRL, or DLA. Table 1. Pertinent Properties of Procured Category A Conventional Fuels Fuel Flash Point, C Viscosity, mm 2 s -1 (cst) ( at -20 C) Aromatics, % (vol) Desired Actual Desired Actual Desired Actual A-1, POSF A-2, POSF A-3, POSF ± ± ±

4 B. Specification Properties Specification properties were obtained from the Air Force Petroleum Agency laboratory at WPAFB, the University of Dayton s laboratories at WPAFB, and Southwest Research Institute in San Antonio TX. Not surprisingly, the fuel properties cannot be varied independently; rather they are interdependent. For example, distilling off portions of a fuel to change the flash point or the freeze point also affects the distillation curve (obviously) and also affects the viscosity significantly. The heat of combustion is directly related to the fuel hydrogen content and properties of most (petroleum-derived) fuels in general can be correlated to the fuel density and average molecular weight/boiling range [6]. A complete set of specification properties was obtained for the Category A fuels. A summary of key properties of the fuels is given in Figure 1, which shows that the three fuels do indeed encompass a wide range of properties within the jet fuel experience base. Tabular data is shown in Appendix A. All three fuels are relatively wide-boiling middle distillate fuels. This characteristic is important to note since narrow-boiling fuels will be a target of research through Category C test fuels described below. The boiling ranges of the three fuels are illustrated in Figure 2. The D86 limits for jet fuel (ASTM D1655) and kerosene (ASTM D6399) are T10 < 205 C and final boiling point <300 C. Note that jet fuel, kerosene, and diesel fuel are all middle distillates per ASTM D4175, but kerosene and jet fuel fall into the class of light middle distillates. Diesel fuel (per ASTM D975) also uses D86, but has 282 C <T90 <338 C, thus has a significantly higher boiling point at the end of the distillation curve. The combustion-related specification properties are shown in Table 2. For reference, the effect on the boiling range of distilling off the light ends of the A-3 fuel to raise its flash point to 70 C is shown in Figure 3. Raising the flash point from 60 to 70 C also increases the viscosity at -20 C from 6.5 to 6.8 cst and raises the freeze point from - 50 C to -49 C. This increased viscosity/freeze point is probably the explanation why the A-3 fuel in Figure 2 has had some of the higher-boiling material removed and has a lower final boiling point than the average fuel the higherboiling materials has been removed to meet the JP-5 freeze point and viscosity requirements. Comparison of various specification distillation methods (ASTM D86, D7345, and D2887) for the three Category A fuels is presented in Appendix C. D86 data can be converted to true boiling point data using equation 3.14 in Reference 6. As shown in Appendix C, ASTM D86 does not represent the true boiling point, but has been in use since the 1930s to conveniently characterize petroleum fractions like jet fuel. Reference 6 defines narrow boiling [petroleum] fraction as one whose ASTM 10-90% distillation slope is < 0.8 C/%. As discussed later, jet fuels are typically right at this limit, so can be defined as narrow boiling with caution. As mentioned previously, the Category A fuels do encompass the wide variety of jet fuels produced (as desired). For example, the range of densities seen in the PQIS data base is shown in Figure 4, with the A-1 and A-3 fuels well out on the ends of the distribution as desired. Density was not a criteria for the Category A fuels, but the viscosity and aromatic requirements also effectively drove the density of the fuels to the edges of the distribution. Figure 1. Property Range of the Category A Conventional Fuels with Respect to Allowed Limits (Red Established, Yellow Proposed) 4

5 Table 2 Specification Test Results for Category A Fuels (from AF Petroleum Agency unless otherwise noted) Category A Fuels Property Test Method Spec limits A-1, A-2, A-3, PQIS 2012 wt mean Density D Flash point, C D93 > Viscosity, -20 C D445 < (cst) Aromatics, vol% D1319 < Heat of D4809 > Combustion, MJ/kg Heat of D Combustion, MJ/kg (SwRI[32]) H content, mass% D3701 > n/a (meas) SwRI H content, mass % D7171 > , n/a (meas) H content, mass % GCxGC > , n/a (UDRI) H/C ratio (based on calculation n/a n/a D3701) Molecular formula GCxGC n/a C 10.8H 21.8 C 11.4H 22.1 C 11.9H 22.6 n/a Derived cetane #, D6890 n/a n/a SwRI Distillation, C D86 IBP * 10% < * 20% * 50% * 90% * FBP < * Engine cetane, D613 n/a n/a SwRI Smoke pt, mm D1322 > Freeze pt, C D5972 >-47 (JP-8) *D86 data from World Survey, since PQIS is a combination of D86 and D2887 5

6 Temperature, C A AF A AF A AF A SwRI A SwRI A SwRI World Survey avg D86 % Distilled Figure 2. ASTM D86 distillation for Category A fuels [SwRI] POSF JP-5, 60 C flash POSF JP-5, 70 C flash Temperature, C D86 % Distilled Figure 3. Change in D86 results obtained by distilling off low boiling point material to raise flash point of A-3 fuel from 60 to 70 C. 6

7 POSF Number of samples POSF POSF Density Figure 4. Density histogram from 2013 PQIS, with Category A fuels labeled. C. Composition The fuel specification properties don t control the fuel composition directly, aside from the 25 vol% limit on total aromatics by ASTM D1319. The indirect effect of the specification limits lead to the distribution of hydrocarbons shown in Figure 5 (A-2/POSF average Jet A) and Figure 6 (hydrocarbon distribution of 55 World Survey fuels averaged together). This data comes from GCxGC measurements by UDRI [10]. The lack of hydrocarbons below about C8 is due to the flash point limit (>38 C). The lack of hydrocarbons above about C17 is due to the freeze point limit (<-40 C for Jet A) and the D86 end point limit (<300 C) (and influenced by the -20 C viscosity limit). There are some inter-relations one can distill off the light ends of a fuel to raise the flash point, but that also tends to increase the viscosity at low temperatures, as mentioned earlier. In any case, typical jet fuels (Figures 5 and 6) have the four types/classes of hydrocarbons (olefins are low in jet fuels) distributed across many carbon numbers. The numerical GCxGC data is included in Appendix A. One area where GCxGC is weak is differentiating the level of branching in iso-paraffins and in side-chains on aromatics and cycloparaffins which can affect combustion properties such as cetane number/ignition delay. NMR is good option for this (e.g., Reference 35 for diesel), but NMR data is not yet available for the Category A fuels. NMR is being used in the development of surrogate fuels, as discussed in Section VI below. Reference 30 includes composition estimates for a number of boiling fractions. The aromatics are the only hydrocarbon class that are not distributed relatively even across the distillation range the aromatics tend to be concentrated in the lower-boiling fractions [30]. This data is used to define the compositional experience base for current/conventional fuels important for evaluating some of the alternative and Category C fuels, whose compositions can be noticeably different from Figure 5 and 6. These figures do confirm that the A-2 fuel is very typical in terms of composition. The composition of the A-3 and A-1 fuels are shown in Figure 7 and 8, respectively. 7

8 6 5 n-paraffins iso-paraffins aromatics cycloparaffins Composition, mass% Carbon number Figure 5. Nominal Category A fuel (A-2) composition 8 7 n-paraffins iso-paraffins aromatics cycloparaffins Composition, mass % Carbon Number Figure 6. Averaged composition of 55 World Survey fuels (Stoddard solvent removed) 8

9 10 8 n-paraffins iso-paraffins aromatics cycloparaffins Composition, mass% Carbon number Figure 7. A-3 composition 10 8 n-paraffins iso-paraffins aromatics cycloparaffins Composition, mass% Carbon number Figure 8. A-1 composition 9

10 D. Fit-for-Purpose Properties The alternative fuel approval process (ASTM D4054) defines a class of properties that are not limited by the specification, but are limited to an expected range for fuels that are acceptable ( fit-for-purpose ). A set of combustion-relevant physical properties was obtained for the A-1, A-2, and A-3 fuels: density vs T, viscosity vs T, Cp vs T, surface tension vs T, vapor pressure vs T. Each fit-for-purpose property is discussed separately below. Published estimation methods are also presented for non-specification properties. 1. Density vs T The density-vs-temperature line (Figure 9) for the A-1 and A-3 fuels are very consistent with the World Survey minimums and maximums. The A-2 fuel density and viscosity is close to the average as reflected in the CRC Aviation Fuel Properties Handbook [3]. Extensive density data on the Category A fuels as a function of pressure has recently been published by NIST [31]; the lowest pressure data (0.5 MPa) is consistent with the data in Figure 9. Other high pressure density data [32] is not entirely consistent with Reference 31. Density is a linear function of temperature in the range shown, which is significantly below the jet fuel critical temperature (~400 C [33]). This is perhaps not surprising since the density of various pure hydrocarbons and petroleum distillates as a function of temperature has been shown to have a similar slope versus temperature [34]. For the temperature range shown, it would seem reasonably accurate to extrapolate density of an unknown jet fuel based on this slope and the 15 C specification density Spec limits at 15 C CRC World Fuel Survey (max) DLA DLA DLA spec spec spec Density, g/cm CRC Handbook Jet A 0.76 CRC World Fuel Survey (min) Temperature, C Figure 9. Density vs T data for Category A fuels 10

11 2. Viscosity vs T The kinematic viscosity-vs-temperature line (Figure 10, note strong temperature dependence) shows that the viscosity of A-1 and A-3 is similar to the minimum and maximum from the World Survey. A-2 is very similar in viscosity behavior to the average shown in the CRC Handbook. Given the strong temperature dependence of viscosity, measurements at -20 C (at least) appear to be warranted for any new fuel, rather than an estimate based on ambient temperature viscosity or based on correlations with other properties. ASTM D7566 has started using a 12 cst limit at -40 C for alternative fuels. The data in Figure 10 has been linearized per ASTM D341 using the correlation: log log (viscosity + 0.7) = A B Log T, where A and B are constants World Survey maximum POSF (A-1) POSF (A-2) POSF (A-3) 5 Viscosity, cst 3 2 World Survey minimum CRC Jet A Temperature, C Figure 10. Viscosity vs T data for Category A fuels. (1 cst = m 2 /s) 11

12 3. Other Combustion-Related Properties (not in specification) A number of other liquid fuel fit-for-purpose properties are relevant to combustion for example, heat capacity, surface tension, vapor pressure, and thermal conductivity. Measured values of these properties were obtained from SwRI for the Category A fuels. Calculated values are also available using petroleum industry correlations dating back many years [6-8]. These properties are usually calculated as a function of density/specific gravity, average boiling point, and/or equivalent molecular weight. The mean average boiling point is typical used for property correlations, although Riazi [6] recommends volume average boiling point for C p. Using the data for the A-2 fuel from Table 2, some typical calculations will be shown. Some of these correlations are very old apologies in advance for the engineering units! API gravity is often used in some of these correlations, where API gravity = (141.5/specific gravity) Specific gravity (SG) = in Table 2 for the A-2 fuel gives API gravity = ) Calculate volume average boiling point: (T10+T50+T90)/3 = 208 C (406 F). Note: Riazi [6] states mean average boiling point ~ T50 for narrow fractions like jet (T50=205 C = 400 F). Charts in Maxwell [8] pg give conversions from volume average boiling temperature to molal, mean average, and weight-average boiling point. These results are molal average = 396 F, mean average = 400 F, weight average = 404 F. 2) ASTM slope = (T90-T10)/(90%-10%) = 0.85 C/% = 1.5 F/% (small slope indicates small correction in going from volume average boiling point to molal average boiling point) (unlike crude) 3) Chart in Maxwell [8] gives a characterization factor of (or equation Kw=T b(in R) 0.33 /specific gravity). Characterization factor is related to paraffinicity of fuel. 4) Fig 5-5 in Nelson [7] yields MW~ 160 (Maxwell [8] ~160); GCxGC ~ 159 (Table 2). Princeton measurements~148. 5) Check Riazi [6] has MW=1.6604X10-4 T b SG (T b in K) ~ 159 6) Dryer et al have patented a direct physical measurement of equivalent molecular weight (patent 9,410,876, August 9, 2016). Comparison ongoing should use direct measurement if available. 7) Nelson [7] chart for heat of vaporization (HOV) ~115 BTU/lb (267 kj/kg); Check CRC gives HOV ~ 275 kj/kg at 208 C (118 BTU/lb) using 208 C as the equivalent normal boiling point (nbp) of the A-2 fuel. Riazi equation for HOV (pg 327, below) ~ 258 J/g at 208 C. Roughly consistent. H vap (nbp)= (T b )(SG ) Lefebvre [9] has H vap (nbp)=( t[in K])/SG, yielding a HOV value of 307 kj/kg a bit higher than the other results. 8) Riazi has equations for T c, P c 737 F (392 C), 21.2 bar result. Maxwell pg 72 eq has 725F pseudocritical T, true T c ~ 735 F. T c data for jet fuels is typically F [33]. References 6-8 can now be used to generate calculated values of heat capacity and surface tension to compare to measured data. 12

13 4. Heat Capacity (C p) vs T Heat capacity (C p) is not a specification property (in contrast to density and viscosity), but is a fit-for-purpose property and was measured at SwRI for the three Category A fuels using ASTM E1269. The results are shown in Figure 11. There is some spread in the data, similar to the spread shown in the CRC Handbook [3], which shows differences among fuels, probably due to density differences. C p calculations based on density and average boiling point [6-8] for the A-2 fuel match the data well. But the density difference between A-1 and A-3 in the calculations is not as large as in the data. Riazi [6] recommends the use of the Kesler/Lee equation: C p L =a(b+ct), with T in K and C p in kj/kg a= kw where Kw is Watson K factor b= SG c= SG This equation leads to C p at 15 C/60 F of 2.06 kj/kg (A-1), 1.96 (A-2), and 1.94 (A-3). Maxwell [8] has a chart which yields C p=1.99 kj/kg-k for A-2 (and 2.15 for A-1, 1.92 for A-3) at 15 C. Lefebvre [9] includes an equation C p=(0.76+0,00335t)/sg 0.5, which yields a Cp for A-2 of 1.92 kj/kg-k at 15 C, pretty consistent with Kesler/Lee equation. Thus, Maxwell s charts show a (calculated) spread in C p across the Category A fuels similar to the SwRI data, while the Kesler/Lee equation predicts significantly less difference. The CRC Handbook yields a C p for typical jet fuels of 1.92 kj/kg-k at 15 C. The World Survey has C p data, but as is discussed in an Appendix in that report [3], there are some issues with the data. Thus, the magnitude of the heat capacity is fairly consistent among the various sources, so the use of the C p data for A-2 for a typical jet fuel seems valid. Note that the data for jet fuels presented in Reference 39 is non-linear with temperature and is inconsistent with the current data. Heat Capacity (Cp), kj/kg-k POSF 10264, A-1 (light JP-8) POSF 10325, A-2 (avg Jet A) POSF 10289, A-3 (JP-5) CRC JP-5 CRC Jet A/A-1/JP-8 CRC JPTS Calculated [6] Temperature, C Figure 11. Heat Capacity vs T data for Category A fuels (1 kj/kg-k = BTU/lb-F) 13

14 The heat capacity can also be calculated from enthalpy data since C p is the slope of enthalpy-vs-temperature at a given temperature. There is some enthalpy data in the literature that can be used to estimate C p [37]. The kerosine in Reference 37 is very similar to the A-2 fuel, as shown in Figure 12 where the D86 data is compared. The enthalpy from Reference 37 is shown in Figure 13 for two pressures. The heat capacity at 100 C can be estimated by the enthalpy change from 24 C to 200 C, divided by the temperature rise. The result is 2.38 kj/kg-k very consistent with Figure 11. Obviously, as the temperature rises above 200 C, fuel vaporization complicates the picture. Incidentally, Lenoir and Hipkin in Reference 37 use the measured enthalpy data for the kerosene to estimate the critical condition for this fuel their tabulated T c~739 F (393 C) and P c~361 psia (24.6 atm) are very consistent with the T c values calculated above from equations in Riazi [6]. Reference 37 has enthalpy data at a number of pressures this data could represent an average jet fuel, given its resemblance to the A-2 fuel. From Figure 13, the heat of vaporization can be estimated as the enthalpy difference between the fuel initially vaporizing (at ~ 225 C) and completing vaporization (at above 250 C), which yields a value of approximately 300 kj/kg roughly consistent with the results presented in section 3 above. Lefebvre [9] shows an enthalpy curve for Jet A that is more detailed than Figure 13, but yields a heat of vaporization of similar magnitude A-2 (POSF 10325) Kerosine [37] 240 Temperature, C D86 % distilled Figure 12. Kerosine D86 data from Reference 37 14

15 atm (1400 psia) 2 atm (30 psia) Ref Enthalpy kj/kg Temperature, C Figure 13. Enthalpy data for kerosine from Reference

16 5. Surface tension vs T Similarly to C p, surface tension (Figure 14) is also not a specification property. In this case, however, the current data from SwRI using ASTM D1331A (du Noüy ring) does not match well with the CRC Handbook (lower than data) or with the calculation (higher than data) although the trends are the same, again the trends are probably due to density differences. There is surface tension data in the World Survey (higher than SwRI data) to get a feel for the magnitude of the density effect, the World Survey data and the current data for A-2 at 22 C is plotted as a function of density in Figure 15. In this case, the measured data is below the World Survey and the calculations are above, but the trend with density is correct. Given the disparity of the literature data and the small span of surface tensions covered by the three Category A fuels, it would seem prudent not to try and draw any conclusion about surface tension effect from the use of the various Category A fuels. For the calculations, Riazi [6] (pg 359) has a surface tension equation as a function of reduced temperature (T r = T/T c) and Kw (Watson K factor). Again, this equation yields trends that agree with the data, but are as far above the data as the CRC results are below the data. Lefebvre [9] shows a graph of surface tension versus T that is consistent with the CRC Handbook data Surface tension, dyne/cm A calc A calc A calc CRC JPTS CRC Jet A/Jet A-1/JP-8 CRC JP Temperature, C Figure 14. Surface tension vs T data for Category A fuels. 16

17 29 28 SwRI measurements A-2 Calculations [6] World Survey all fuels 22 C Surface tension, dyne/cm Density, g/cm 3 Figure 15. Surface tension vs density data (22 C) for various fuels. 17

18 6. Vapor Pressure vs T Vapor pressure is not a specification property. Vapor pressure is not independent of the other properties discussed it is likely that one could calculate the vapor pressure from the D86 and/or flash point. For example, the CRC Handbook shows separate vapor pressure curves for JP-8/JetA/Jet A-1 (flash >38 C) and JP-5 (flash > 60 C). The vapor pressures of the three Category A fuels was measured at SwRI yielding the results shown in Figure 16. And, yes, the data comes as vapor pressure in psia versus temperature in C, so those mixed units are plotted directly. The vapor pressures track with D86 and flash point, as expected JP Jet A JP-5, psia Vapor pressure, psia 3 2 CRC JP-8 CRC JP-5 Range in HEFA Research Report Temperature, C Figure 16. Vapor pressure vs T data for Category A fuels 18

19 7. Thermal conductivity vs T Thermal conductivity data was not obtained for the Category A fuels. Thermal conductivity of liquid jet fuel is not a specification property, and is presented in the CRC Handbook as a single line for all jet fuels (but a line that has moved over the years the line from the 2004 Handbook best matches the data collected in the 1980s and 1990s). It has not been typically measured for alternative fuels due to its expected small variation with fuel composition. Riazi [6] presents an equation for thermal conductivity that basically splits the difference between the two CRC lines (Figure 17) when the values for the A-2 fuel are plugged into the equation. Where T b and T are in K and k (thermal conductivity) is in W/m-K. Lefebvre [9] includes a simpler equation: k=( )/sg, which yields a thermal conductivity of for A-2 at 15 C, well above any of the data in Figure 17. In the absence of other data, approximating the thermal conductivity of a jet fuel with the data from the CRC 2004 (or later) handbook is probably the best option Thermal conductivity, W/m-K CRC jet 2004 CRC jet 1983 Riazi equation for A Temperature, C Figure 17. Thermal conductivity as a function of temperature 19

20 8. Heat of formation For some computer calculations, the heat of formation of various fuels is needed to calculate flame properties. As described in Reference 38, the heat of formation can be (back)calculated from the measured hydrogen content and measured heat of combustion. CaHb + (a+ (b/4)) O2 a CO2 + (b/2) H2O a(δhf CO2) + (b/2)(δhf H2O (gas)) - (ΔHf fuel) (a+(b/4))(δhf O2) = Heat of combustion, in (typically) kcal/mole fuel, ΔHf where the only unknown in the second equation is the heat of formation of the fuel (since b/a=h/c and the heat of formation of water and CO 2 are known). In this calculation below, the fuel H/C ratio is used to artificially define the fuel as (e.g.) CH This abstraction is used to initially (mis)define a mole of fuel to end up with the heat of formation in cal/g (as is typical in some calculations). Since a mole of fuel, or the fuel equivalent molecular weight, is an abstraction typically heat of formation is reported in cal/g. However, an equivalent molecular weight by GCxGC or other means can be used to get heat of formation in terms of kcal/mol that is more realistic than defining the fuel as CH Such a calculation is performed in Table 3 below. Within the accuracies of the measured H content and heat of combustion, it appears the heat of formation of the three Category A fuels is essentially the same. Table 3. Heat of formation calculations Fuel wt% H H/C molar Mass heat of Mass heat Heat of Heat of (meas) (calc from H comb, MJ/kg of comb, formation MW, formation, SwRI content) (SwRI, meas kcal/g (calc), cal/g GCxGC kcal/mol D3701 D4809) A A A calorie/gram = Joules/gram = 1.8 BTU/lb 20

21 III. Test Fuels Category C A. Overview of Category C Fuels [1] The main objective of the Category C test fuels was to identify hydrocarbon blends that had unusual (outside of experience) properties, such as narrowly distributed aromatics at the front end of the boiling range or high viscosities. The CRATCAF Phase IIa report [2] had a list of potential blending components and several blended fuels listed for testing in later phases. For the NJFCP, in consultation with the OEMs, six test fuels of interest were selected based on properties and availability; components were acquired in neat form as necessary and blending was performed at AFRL. It was decided that these test fuels were to be tested in their neat form without blending with petroleum jet fuels to increase the likelihood of observing differences. Thus some of the Category C fuels do not meet all the jet fuel specifications (like density), so they are best called test fuels rather than jet fuels. Six Category C test fuels were chosen initially for study in Year 1. One of them (C-6, a high-cycloparaffin test fuel) has not been obtained in sufficient quantities for testing and could not be included in this paper. The characteristics of the remaining five fuels are compared in Table 4. A seventh Category C test fuel considered was the high flash point A-3 fuel, with the front end of the fuel distilled off to raise the flash point to 70 C, as described in the prior section. This test fuel is available for testing in subsequent years if there is interest in a higher-flash-point fuel. In the 2015 mid-year meeting of the NJFCP team, results for tests of the Category C test fuels were reviewed and it was decided to focus on the C-1 and C-5 test fuels in the near-term, since the various tests seemed to be the most sensitive to those fuels, and each represented extremes in chemical and physical properties. Combustion test results for the various Category C fuels will be reported separately (e.g., [28,29]). Table 4. Category C Fuel Types NJFCP Test Fuels NJFCP Fuel ID C-1 C-2 C-3 C-4 C-5 POSF numbers 11498, 12368, Composition Notable characteristics Gevo ATJ; C12/C16 highlybranched iso-paraffins Very low cetane, unusual boiling range 11813, % C14 isoparaffins; 16% 1,3,5 trimethyl benzene from Swift* On-spec fuel, extremely chemicallyasymmetric boiling 12341, % A-3; 36% Amyris farnesane (C15 isoparaffin) High viscosity fuel, at -20 C viscosity limit for jet fuel 12344, , 12713, 12789, % Sasol* IPK (C10-C13 highly branched isoparaffins)/40% C-1 Low cetane, conventional, wide-boiling range 74% C10 isoparaffins, 26% 1,3,5 trimethyl benzene Very flat boiling range (fuel boils at one temperature) range *Any identifications of commercial products within this paper is for information only and does not indicate recommendation or endorsement by FAA, AFRL, or DLA. 21

22 B. Specification Properties The specification properties of the various Category C fuels are shown in Table 5. Table 5 Category C Fuel Properties C-1 C-2 C-3 C-4 C-5 Property Test method Density D Flash point, C D Viscosity, -20 D , C (cst) Aromatics, D vol% Heat of D4809 AF, SwRI 43.88± , , , , 43.0 Combustion, MJ/kg (15 meas.) H content, D3701 SwRI 15.28, mass% (meas) H content, D7171 AF mass % (meas) H content, D5291 SwRI mass % (meas) H content, GCxGC mass% H/C ratio calculation (based on D3701) Molecular GCxGC C 12.6H 27.2 C 12.4H 24.8 C 12.8H 25.3 C 11.4H 24.8 C 9.7H 18.7 formula Derived D (range cetane # ) Smoke pt, D mm Freeze pt, C D5972 < < Distillation, D86 C IBP % % % % FBP Some of the differences among the Category C test fuels are evident in Figure 18, where the various distillation curves are plotted (compare to Figure 2 for the petroleum-derived/conventional fuels). C-5 was set up to be a fullyformulated fuel (including aromatics), but has an extremely flat boiling range (i.e., boiling at essentially one temperature, like a pure hydrocarbon fluid). This test fuel was created by blending 1,3,5 trimethyl benzene with a C 10 iso-paraffinic solvent, both of which boil at roughly 165 C. This test fuel was designed to evaluate the impact of a very limited vaporization range of the fuel on combustor operability. However, this fuel also has a low (outside of experience) viscosity which may impact its performance and make interpretation more difficult. In contrast, the C-2 22

23 test fuel is a fully formulated fuel, but has the same low-boiling aromatic compound combined with higher-boiling C 14 iso-paraffins. This fuel is thus the opposite of the C-5 fuel in one sense with one class of materials vaporizing at an entirely different temperature than the other. Temperature, C C AF C AF C AF C AF C AF C SwRI C SwRI C SwrI C SwRI C SwRI "low cetane bimodal" "high viscosity" bimodal "low cetane wide boiling" 160 "flat" D86 % Distilled Figure 18. D86 distillation curves for Category C fuels The C-1 test fuel is representative of a fuel composed of heavily branched iso-alkanes. This specific fuel is notable for having two carbons numbers only (12 and 16) and an extremely low derived cetane number (16) relative to other fuels (typical jet fuels have DCNs of 40-50). Testing of the neat C-1 fuel in this program is to determine the effect of low cetane on combustor operability. In an attempt to isolate the effect of low cetane number from the unusual carbon number distribution, the C-4 fuel was created by blending C-1 with a C 9 to C 12 blend of isoparaffins (Sasol IPK with a derived cetane number of 31) to create a test fuel with more typical boiling characteristics, but with an intermediate (but still low) cetane number of 28. The C-3 test fuel was formulated to have a viscosity at the jet fuel specification limit (8 cst or mm 2 s -1 at -20 C). This test fuel was created by adding farnesane (trimethyl dodecane) to the A-3 conventional high-viscosity JP-5 fuel, with the effect of farnesane addition shown in Figure 19. Note that the C-3 fuel exceeds the 12 cst at -40 C limit used in ASTM D7566 this 12 cst limit has been proposed to replace the 8 cst limit at -20 C to ensure proper APU altitude start after cold soak. 23

24 8.2 POSF = 8 cst at -20 C 8 Viscosity at -20 C, cst Vol % farnesane in JP (A-3) Figure 19. Increase in viscosity by adding farnesane (2,6,10 trimethyl dodecane) to JP-5 (A-3). C. Composition The composition of the Category C fuels is simpler in general than that of the Category A fuels. The two simplest Category C fuels are C-2 (Figure 20) and C-5 (Figure 21). The C-5 fuel was designed to be a fully-formulated fuel that meets all of the jet fuel specification requirements, but has a very flat boiling range (boils at one temperature). In contrast, the C-2 fuel was designed to be a fully-formulated fuel that had a bimodal boiling distribution, with an aromatic component boiling first, followed by an iso-paraffin. It was determined that isomerization of a C14 alphaolefin to the point that it meets the jet fuel freeze point required essentially complete removal of n-c14. As shown in Figure 20, the C-2 fuel is predominantly C14 iso-paraffins blended with 1,3,5 trimethyl benzene (C9). The thought behind this fuel was that preferential vaporization (if important) of the trimethyl benzene would create a very difficult fuel to ignite/burn. In contrast, the C-5 fuel would not have preferential vaporization issues, but might evaporate quite differently from a conventional fuel. The C-5 fuel s composition is shown in Figure 21. Figure 18 shows that this fuel is indeed quite flat boiling. Properties of these fuels are shown in Appendix B. The C-3 fuel is a modification of the A-3 (JP-5) fuel, with its viscosity increased by adding farnesane (2,6,10 trimethyl dodecane) to hit the specification limit (8 cst at -20 C), as discussed in reference to Figure 19. The resulting composition is shown in Figure 22. The C-1 fuel is was designed to be the lowest cetane jet fuel available, which turned out to the Gevo ATJ fuel, consisting primarily of C12 and C16 iso-paraffins. There was some concern about the bimodal nature of the D86 curve for this fuel (C-1), so the C-4 fuel was created by blending in Sasol IPK (another relatively low-cetane fuel, but one with a very different boiling range. The resulting composition is shown in Figure 23. This fuel has a cetane number in the mid 20s, but a much broader boiling range. 24

25 80 70 n-paraffins iso-paraffins 1,3,5 trimethyl benzene Composition, mass% Carbon number Figure 20. C-2 fuel composition n-parafffins iso-paraffins 1,3,5 trimethyl benzene Composition, mass% Carbon number Figure 21. C-5 fuel composition 25

26 10 8 n-paraffins iso-paraffins aromatics cyclo-paraffins 36 Composition, mass% Carbon number Figure 22. C-3 fuel composition (A-3 + farnesane to increase viscosity) 80 Isoparaffins, mass% Gevo ATJ, POSF 11498, "C-1" Sasol IPK, POSF 7629, "CRATCAF C1" POSF 12344, "C-4", 60/ /11498 blend Carbon number Figure 23. C-1 and C-4 fuel compositions. 26

27 D. Fit for Purpose Properties Fit-for-properties were measured for the Category C fuels, as for the Category A fuels. The correlations discussed in the previous section were validated for petroleum fractions, so their relevance to these constructed Category C fuels is questionable so calculations were not performed for the Category C fuels. In Figure 24, density as a function of temperature is shown (again) to be linear with temperature over the range of -40 to +85 C. Two sets of measurements are represented SwRI (5 to 85 C) and UDRI (-20 and -40 C). For Category C fuels that are blends, the density was linear with blending, as shown in Figure 25. Thus, even for very asymmetric mixtures like C-2, blending of these hydrocarbons was linear in density. C-1, C-4, and C-5 do not meet the jet fuel density spec at 15 C ( ). Viscosity versus temperature for the Category C fuels is shown in Figure 26. Most blends fall within experience, except C-5 (low) and C-3 (high). C-3 was designed to be outside of experience in viscosity this is illustrated in Figure 26. Alas, C-5 was designed to be flat-boiling (which it is), but the available ingredients led to a fuel with an unusually low viscosity, as shown in Figure 26, clouding the combustion results to some extent. 0.9 Density, g/cm specification range A A A , C , C C , C , C Temperature, C Figure 24. Density data for Category C fuels (note - points for C-1 lie underneath C-4 fuel symbols) 27

28 Density in 12223; TMB in C14 Density in 12345; TMB in C10 Density in 12341; farnesane in JP Density, g/cm Vol% TMB or farnesane Figure 25. Demonstration of linear blending in C-2 (POSF 12223), C-5 (POSF 12345), and C-3 (POSF 12341). 15 D445 Viscosity, cst 10 5 JP-5/high visc Jet A/Jet A-1 A A A C C C C C WS min 720 WS max Temperature, C Figure 26. Temperature dependence of viscosity for Category C fuels 28

29 Figure 27. Viscosity (-20 C) for several Category C fuels that are outside of experience. Heat capacity (specific heat) data for the Category C test fuels is shown in Figure 28. Many properties seem to trend with density heat capacity in this case is not one of them (15 C densities are shown in the legend). It isn t clear if the differences among the fuels is real or an artifact of the measurement. The CRC Handbook shows separate C p lines for JP-5, Jet A/Jet A-1/JP-8, JP-TS, and JP-4/Jet B, varying by about 0.1 kj/kg-k at a given temperature. The CRC Handbook shows a C p for Jet A/Jet A-1/JP-8 of 2.3 kj/kg-k at 100 C, consistent with the A- 2 value shown in Figure 28. A recent analysis of specific heat data for jet fuels, alternative fuels, and pure jet fuelrange hydrocarbons [40] has shown that heat capacity trends more with chemical composition than density. Pure n- paraffins and iso-paraffins lie above the CRC Handbook jet fuel line, while aromatics lie below. That is broadly consistent with the data of Figure 28, although the low C p value for the C-1 fuel (a pure iso-paraffin) is puzzling, as are the relatively high C p values for the C-3 fuel (11% aromatics). Figure 29 shows the surface tension data for the Category C fuels. The data is broadly consistent with the Category A fuels, although trends with fuel properties are not clear. As shown in Figure 30, surface tension at a given temperature correlates roughly with density, although there is significant scatter. In Figure 31, vapor pressure for the Category C test fuels is generally as expected, with the high-viscosity C-3 fuel (based on JP-5) at the low end and the C-5 fuel on the high end (as expected from its composition). 29

30 Heat Capacity (Cp), kj/kg-k POSF 11498, C-1, POSF 12223, C-2, POSF 12341, C-3, POSF 12344, C-4, POSF 12345, C-5, POSF 10325, A-2, Temperature, C Figure 28. Heat capacity as a function of temperature for Category C fuels (SwRI). 15 C density is shown in legend. 30 Surface tension, dyne/cm A A A C C C C C Temperature, C Figure 29. Surface tension as a function of temperature for Category C fuels (SwRI). Compare Figure 14 for CRC Handbook data. 30

31 30-10 C 22 C 28 Surface tension, dyne/cm C-4 C-1 C-5 C-2 A-1 C-3 A-2 A Density (15 C) g/cm 3 Figure 30 Surface tension correlation with density 5 C C C C Vapor pressure, psia 3 2 C CRC JP-8 CRC JP-5 Range in HEFA Research Report Temperature, C Figure 31 Vapor pressure data for Category C fuels 31

32 As part of this program, blends of the A-2 Jet A and the C-1 test fuel were made see Figures 32 to 35. Bulk properties such as heat of combustion and hydrogen content should be linear with blending the non-linearity in the H content is due to scatter in the data/measurement. Density is expected to be linear with blending (as mentioned earlier) and is (Figure 34). Viscosity was not expected to be linear with blending and was not (especially at lower temperatures Figure 35). Over the course of four years, multiple batches of the C-1 fuel were received and multiple measurements were made of the properties. From the AF Petroleum Agency lab, the heat of combustion measurements were 43.93, 43.90, 43.94, 43.95, 43.77, 44.00, 43.85, 43.74, 43.86, and MJ/kg. Measurements at SwRI were and MJ/kg. So, the mean is and the standard deviation is 0.09 MJ/kg. The H content by D7171 (AF) and D3701 (SwRI) was 15.4, 15.7, 15.2, 15.4, 15.5, 15.2, 15.5, 15.2, 15.2, 15.5, and (D3701). The mean was thus 15.4 mass% H, and the standard deviation was 0.17 mass%. Since the C-1 fuel is mostly pentamethyl heptane (C12H26) isomers, the calculated H content can be estimated as 26(1.008)/(26(1.008)+12(12.011))~15.38 mass%. So the mean value on H content is apparently relatively accurate, but the standard deviation is higher than desirable. ASTM D5291 gives mass% hydrogen. 44 y = x R= Heat of combustion, MJ/kg Vol% C-1 in A-2/C-1 blend Figure 32 Heat of combustion as a function of blend ratio for C-1/A-2 blends 32

33 16 y = x R= H content, mass % (D7171) Vol% C-1 in A-2/C-1 blend Figure 33 H content as a function of blend ratio for C-1/A-2 blends 0.81 y = x R= Density, g/ml Vol %C-1 in A-2/C-1 blend Figure 34 Density as a function of blend ratio for C-1/A-2 blends 33

34 Viscosity at 40 C, cst Viscosity at -20 C, cst A-2 %C-1 in A-2/C-1 blend C-1 Figure 35. Viscosity as a function of blend ratio for C-1/A-2 blends It was not part of this program, but several sets of blending data for the C-1 (Gevo ATJ fuel) have been obtained for derived cetane number. As shown in Figure 36, the blending curve of the low cetane C-1 fuel is not linear with a number of fuels. A number of different samples of the C-1 fuel have been tested, with DCN ranging from 15.1 to ASTM D6890 Derived Cetane Number JP-8 "low cetane" JP-5 JP JP Volume % ATJ Figure 36. Derived cetane number for various blends of C-1 fuel with jet fuel 34

35 Heat of formation As mentioned above, the heat of formation might be needed for the Category C fuels, and can be calculated from the measured hydrogen content and heat of combustion (Table 6). There is more variation in heat of formation amongst the Category C fuels than the Category A fuels. Fuel wt% H (meas) SwRI D3701 Table 6 Heat of formation calculations H/C molar Mass heat of Mass heat (calc from H comb, MJ/kg of comb, content) (AF/SwRI, kcal/g meas D4809) Heat of formation (calc), cal/g MW, GCxGC Heat of formation, kcal/mol C C C C C

36 s IV. Other Average Jet Fuels Distributed A number of previous average fuels have been distributed to researchers over the past 15 years. POSF 6169, 5699, 4177, and 3773 are JP-8 fuels obtained from the active airfield at WPAFB. It is perhaps coincidental that they are fairly average. In contrast, POSF 4658 was created by taking 2000 gallons of each of five different Jet As (each from a different part of the U.S.) and blending them all together. As can be seen in Table 7, all of these fuels are similar to the A-2 fuel and are, indeed, pretty average in terms of their properties. GCxGC data is presented in Appendix D. Table 7 Specification data for previous reference fuels Property Test Method Spec limits POSF 4658 POSF 6169 POSF 5699 POSF 4177 POSF 3773 PQIS 2012 wt mean Density D Flash point, C D93 > Viscosity, -20 D445 < C (cst) Aromatics, D1319 < vol% Heat of Combustion, MJ/kg D4809 > (D3338) H content, mass% (calc) D3343 > Derived cetane number D n/a n/a Smoke pt, mm > , , Freeze pt, C >-47 (JP- 8) GCxGC est formula C11.7H22.6 C11.3H

37 V. Alternative Jet Fuels A. Introduction Large alternative fuel programs established in 2006 have led to the certification of 5 alternative drop-in hydrocarbon jet fuels in ASTM D7566 by mid The properties of these fuels can be found in Research Reports [11,12,15,18] and other reports [13,14,16,17]. The purpose of this section is to list POSF numbers for common alternative fuels shipped by the Air Force since Supplies of many of these fuels still exist. Often, when fuels are combined, moved, received back from a shipment, or additives are added a new number is assigned for a fuel that is chemically identical to a previous number. For example, Sasol IPK was purchased in a 300,000 gallon batch by the AF. When 6000 gallons of that was received by AFRL, it was given the number POSF A large part of the original 300,000 gallons went to various AF locations for testing when the excess fuel was collected at WPAFB by the AF several years later and combined with remaining POSF 5642 at WPAFB, the IPK was given a new number POSF 7629, although the hydrocarbon composition of the fuel is identical to the original shipment/batch. Typically a set of jet fuel spec tests is associated with a given POSF ID number. Thus, for example, the numerous ATJ batches listed below all have a set of (very similar) spec tests so the consistency of the various batches can be assessed. (a) ASTM D7566 Annex 1 Synthetic Paraffinic Kerosene (aka Fischer-Tropsch fuels) --Sasol IPK (made in South Africa), AF purchase, 2008, same large batch POSF 5642, 7629, 5959, 5654, 7279, Shell GTL (made in Bintulu Malaysia), AF purchase, 2007, same large batch POSF 5172, Syntroleum GTL (made in Tulsa OK), AF/Army purchases 2004: POSF 4734 (drums), 2005: POSF 4820 (drums), 2006: POSF 5018 (tanker trucks). Composition very similar between 2004, 2005, 2006 batches --Rentech GTL POSF 5698, 7457 (note these first three fuels make up most of the data in the SPK Research Report [11-12], although the IPK data in the Research report was largely from Sasol production prior to the batch produced for the Air Force) (b) ASTM D7566 Annex 2 - Hydroprocessed Esters and Fatty Acids (HEFA) (aka Hydroprocessed Renewable Jet (HRJ)) --Tallow-based HRJ fuel produced by UOP for AF in 2010 (large batch) POSF 6308, 6346, 9584, Camelina-based HRJ fuel produced by UOP for AF in 2009 POSF 6152, large batches=posf 11714, 7720 (not identical to 6152) --Mixed-fat-based HRJ produced by Dynamic Fuels/Syntroleum in 2010 (aka R-8 ): large batch POSF 7272, 7635; 2008 drums: POSF (c) ASTM D7566 Annex 5 - Alcohol-to-Jet Synthetic Paraffinic Kerosene Gevo initial batch - POSF 7504 (1 drum, 2011, lower C16/C12 ratio than all other batches, which are basically identical) Subsequent Gevo batches (many were 6000 gallons) - POSF 7695, 7699, 7712, 7788, 7817, 8092, 8158, 8289, 8438, 9641, 10151, 10262, 10373, 11498, 12368, ( ) 37

38 VI. Surrogate Fuels The focus of this section is much narrower than surrogate fuels in general [19-21]. Here we discuss only two jet fuel surrogates produced in the hundreds of gallons to compare experimentally to results for the real fuel which the surrogate is supposed to mimic. This builds on the approach developed in a recent MURI [22, 23]. Technically, the surrogate fuel could be designed to simulate only the combustion chemistry of the parent fuel, but a surrogate which also mimics the physical properties has the potential to be used to model the fuel throughout the entire injection/combustion process. One can easily envision a blend of a few hydrocarbons that would match the bulk physical properties of the fuel being modeled density, H/C ratio (and sooting), average molecular weight, cetane number. The difficulty comes with trying to match properties that are dependent upon the multicomponent nature of the fuel, such as viscosity and (especially) boiling range (as approximated by ASTM D86 distillation). A key simplification applied is that the boiling range of the entire fuel is being simulated the boiling range of each class of hydrocarbons is NOT being simulated that would require many more surrogate components. Some more complex surrogates for jet fuels in diesel engines (4 components) [24] and diesel fuel in diesel engines (8-11 components) [25] have been described. From ongoing testing in NJFCP, it appears that matching the boiling range of individual fractions is not necessary in the turbulent environment of a gas turbine combustor (as noted above and in other papers at this meeting, fuels with abnormal boiling distributions in Category C were generally found to burn similarly to conventional fuels). These initial surrogates were combination of three hydrocarbons an n-paraffin, an iso-paraffin, and an aromatic. The need for including a cyclo-paraffin to match all the major classes of jet fuels is still being debated. In an overall sense, a surrogate could not consist of just n-paraffins, because the cetane number would be too high and the density too low. Inclusion of aromatics increases the density and decreases the cetane number, while the effect of adding isoparaffins would depend upon the degree of branching. Typically, lightly-branched iso-paraffins make up the largest hydrocarbon class in conventional jet fuels and have a relatively high cetane number. Highly branched iso-paraffins have low cetane numbers, as do aromatics. Alas, the flexibility of surrogate creation is limited by the cost of the ingredients (500 gallons was deemed to be a good batch size, enabling in use of the surrogate in several larger rigs). This cost issue was recognized early on [19, 20]. In 2016, with jet fuel cost of $2-3/gallon, the cheapest surrogate ingredients were on the order $50/gallon. Some desirable ingredients were hundreds of dollars/gallon and more. Lightly-branched iso-paraffins in the jet fuel range are especially expensive. Cyclo-paraffins in the jet fuel range are also currently very high-priced, although some solvent options may be available. With that background in mind, two surrogates for the average Jet A fuel described earlier (A-2, POSF 10325) were blended in 2016, as shown in Table 8. The surrogates were 1,3,5 trimethyl benzene and iso-octane blended with either n-dodecane or n-hexadecane to match H/C ratio, smoke point, and DCN [22,23]. Typically technical grades of the hydrocarbons were purchased. These surrogates roughly match the DCN, H/C ratio, and smoke point of the average jet fuel (A-2). The density and average molecular weight of surrogate #1 is somewhat lower than the target jet fuel. Surrogate #2 increases the MW and density by replacing n-dodecane with n-hexadecane (Table 9). The ASTM D86 boiling range is shown in Figure 37, and compared to the three Category A fuels. The lower initial boiling point is due to the compromise of being forced to use iso-octane as the iso-paraffin. Technical grade iso-octane can be found for under $100/gallon, no doubt because of its use as a component of gasoline primary reference fuels. The preference was to have iso-dodecane (penta-methyl heptane) as a component [26] but it was not available in drum quantities at the time of this paper, even at $100/gallon. Lightly-branched iso-paraffins (e.g., 2-methyl decane) cost about $1/gram. As discussed in Section 3, 1,3,5 trimethyl benzene has a boiling point of ~160 C, so the replacement of iso-octane with iso-dodecane would bring the initial boiling point of such a surrogate up into the jet fuel range (IBP ~ 160 C) and increase the MW and density. Note also in Figure 37 that the substitution of dodecane with hexadecane raises the last 50% of the boiling range significantly even above typical jet fuels so an optimized surrogate might not use n-hexadecane. Some of the specification properties of the surrogates are shown in Table 10. The combustion behavior of these two surrogates is being assessed and some results may be available in other papers presented at this meeting [28, 29]. The relationship between the ASTM D86 distillation curve and the actual fuel evaporation under combustor conditions is uncertain. It was surprising that the three-component nature of the surrogates was not more evident in the D86 curve, so two other distillation techniques were utilized to characterize the surrogates. ASTM D7345 mini-distillation closely resembled D86, while the ASTM D2887 simulated distillation (GC) technique did more closely resemble the actual boiling points of the three components, as shown in Figure 37 and 38. The D7345 and D2887 data for the Category A fuels is shown in Appendix C. 38

39 Table 8 Surrogate composition data Surrogate 1 as specified, vol % Surrogate 1 (POSF 12765) as blended - UDRI GCxGC, vol % Surrogate 2 as specified, vol % n-dodecane Surrogate 2 (POSF 12785) as blended - UDRI GCxGC, vol % n-hexadecane iso-octane ,3,5 trimethyl benzene iso-c12 (impurity in dodecane) 1.3 iso-c16 (impurity in hexadecane) 1.5 n-c10 (impurity) 0.3 n-c14 (impurity) Table 9 Surrogate property data (Princeton surrogate calculator) Surrogate 1 Surrogate 2 Target (A-2 (POSF 10325)) Density ASTM D4052, g/cm DCN Smoke point/tsi MW H/C D86 distillation, C Table 10 AFRL surrogate property data on bulk batches Surrogate 1 (POSF 12765) Surrogate 2 (POSF 12785) A-2 (POSF 10325) IBP % % % % FBP ASTM D1319 aromatics, vol% 23 25, 26.8, ASTM D93 flash point, C ASTM D445 viscosity at -20 C, cst 2.7 n/a 4.5 ASTM D7171 H content, mass% , ASTM D4809 heat of comb., MJ/kg ASTM D5972 freeze point, C Molecular formula (GCxGC) C 10.3H 20.1 C 11.2H 21.9 C 11.4H 22.1 ASTM D4052 density, g/ cm

40 ASTM D86 Temperature, C POSF surrogate 1 POSF surrogate 2 POSF Jet A POSF JP-8 POSF JP % Distilled Figure 37. ASTM D86 distillation results for surrogates as compared to conventional fuels D D D2887 Temperature, C vol% n-c12, 18.4% iso-c8, 22.2% TMB b.p. 216C 99 C 165C % distilled Figure 38. ASTM D86/D7345/D2887 distillation results for surrogate fuel #1 40

41 D D D2887 Temperature, C vol% n-c16, 25.1% iso-c8, 22.2% TMB b.p. 287 C 99 C 165C % distilled Figure 39. ASTM D86/D7345/D2887 distillation results for surrogate fuel #2 Note from Table 9 that the freeze point of surrogate 2 is higher than desired for jet fuel, and the viscosity could not be measured at -20 C because of solidification of the n-hexadecane. However, the density and average MW are close to the desired values. A third surrogate is being planned with iso-dodecane in place of iso-octane, which should enable matching DCN, MW, density, smoke point, and viscosity (and perhaps D86 distillation) more effectively than surrogates #1 and #2. 41

42 VII. Conclusion This paper summarizes the various fuels being used as reference fuels for the National Jet Fuel Combustion Program, as well as other fuels used recently as reference fuels. The paper includes the physical properties of the fuels. The fuels cover a wide range of combustion-related properties. The fuels are available to researchers outside of the National Jet Fuel Combustion Program. Acknowledgments DLA funding support for CRATCAF and NJFCP programs is gratefully acknowledged, as is their assistance with fuel procurement and shipping. References [1] Colket, M., et al, An Overview of the National Jet Fuels Combustion Program, AIAA , January [2] Edwards, T., Moses, C., Dryer, F., Evaluation of Combustion Performance of Alternative Aviation Fuels, AIAA , July Also AFRL-RZ-WP-TR and AFRL-RQ-WP-TR (limited distribution). [3] Coordinating Research Council, Handbook of Aviation Fuel Properties, CRC Report 530, 1983/2004. [4] Hadaller, O. J., & Johnson, J. M., World Fuel Sampling Program, Coordinating Research Council Report 647, (often termed the CRC World Fuel Survey) [5] Defense Logistics Agency (DLA), Petroleum Quality Information System (PQIS), annual reports. [6] Riazi, M. R., Characterization and Properties of Petroleum Fractions, ASTM Stock Number: MNL50, ASTM, West Conshohocken, PA, [7] Nelson, W. L., Petroleum Refinery Engineering, McGraw-Hill, New York, [8] Maxwell, J. B., Data Book on Hydrocarbons Application to Process Engineering, Van Nostrand, NY, [9] Lefebvre, A., Gas Turbine Combustion, Hemisphere, NY, Also Lefebvre, A.H., Fuel Effects on Gas Turbine Combustion Ignition, Stability, and Combustion Efficiency, Journal of Engineering for Gas Turbines and Power, Vol. 107, pp , And Lefebvre, A. H., Fuel Effects on Gas Turbine Combustion, AFWAL-TR , Jan [10] Striebich,R. C., Shafer, L. M., Adams, R. K., West, Z. J., DeWitt, M. J., Zabarnick, S., Hydrocarbon Group- Type Analysis of Petroleum-Derived and Synthetic Fuels Using Two-Dimensional Gas Chromatography, Energy & Fuels, Vol. 28, pp , [11] Moses, C., Comparative Evaluation of Semi-Synthetic Jet Fuels, final report for CRC Project No. AV-2-04a, September [12] Moses, C., Comparative Evaluation of Semi-Synthetic Jet Fuels - Addendum: Further Analysis of Hydrocarbons and Trace Materials To Support Dxxxx, final report for CRC Project No. AV-2-04a, April (DXXXX later became D7566) [13] Moses, C. A., Stavinoha, L. L., & Roets, P., Qualification of Sasol Semi-Synthetic Jet A-1 as Commercial Jet Fuel, SwRI Report 8531, November [14] Rahmes, T. F., Kinder, J. D., Henry, T. M., Crenfeldt, G., LeDuc, G. F., Zombanakis, G. P., Abe, Y., Lambert, D. M., Lewis, C., Juenger, J. A., Andac, M. G., Reilly, K. R., Holmgren, J. R., McCall, M. J., & Bozzano, A. G. (2009), Sustainable Bio-Derived Synthetic Paraffinic Kerosene (Bio-SPK) Jet Fuel Flights and Engine Tests Program Results, AIAA , July [15] The Boeing Company, UOP, & United States Air Force Research Laboratory. (2011). Evaluation of Bio- Derived Synthetic Paraffinic Kerosenes (Bio-SPKs). West Conshohocken, PA: ASTM International. [16] Edwards, T., Shafer, L., Klein, J., U.S. Air Force Hydroprocessed Renewable Jet (HRJ) Fuel Research, AFRL-RQ-WP-TR , July [17] Striebich, R., Shafer, L., et al, Dependence of Fuel Properties During Blending of Iso-paraffinic Kerosene and Petroleum-Derived jet Fuel, AFRL-RZ-WP-TR , Nov [18] Edwards, T., Johnston, G., et al, Evaluation of Alcohol to Jet Synthetic Paraffinic Kerosenes (ATJ-SPK), ASTM Research Report, [19] Colket, M. B., et al, "Development of an Experimental Database and Kinetic Models for Surrogate Jet Fuels," AIAA Paper , Jan [20] Colket, M. B., et al, Identification of Target Validation Data for Development of Surrogate Jet Fuels, AIAA Paper , January

43 [21] Edwards, T., Maurice, L. Q., Surrogate Mixtures to Represent Complex Aviation and Rocket Fuels, Journal of Propulsion and Power, Vol. 17(2), pp , [22] Dryer, F., Ju, Y., Sung, C.-J., Brezinsky, K., Santoro, R.J., Litzinger, T., and Curran, H., Generation of Comprehensive Surrogate Kinetic Models and Validation Databases for Simulating Large Molecular Weight Hydrocarbon Fuels, 2007 MURI Topic: Science-Based Design of Fuel-Flexible Chemical Propulsion/Energy Conversion Systems Contract/Grant No. FA July 1, 2007 June 30, [23] Dooley, S., Won, S. H., Chaos, M., Heyne, J., Ju, Y., Dryer, F.L., Kumar, K., Sung, C-J., Wang, H., Oehlschlaeger, M.A., Santoro, R.J., Litzinger, T., A jet fuel surrogate formulated by real fuel properties, Combustion and Flame Vol. 157 No. 12, 2010, pp [24] Kim, D., Martz, J., Violi, A., A surrogate for emulating the physical and chemical properties of conventional jet fuel, Combustion and Flame, Vol. 161, pp , [25] Mueller, C. J., Methodology for Formulating Diesel Surrogate Fuels with Accurate, Compositional, Ignition-Quality, and Volatility Characteristics, Energy & Fuels, Vol. 26, pp , [26] Won, S. H., Haas, F. M., Tekawadeb, A., Kosibab, G., Oehlschlaeger, M., Dooley, S., Dryer, F. L., Combustion characteristics of C4 iso-alkane oligomers :Experimental characterization of iso-dodecane as a jet fuel surrogate component, Combustion and Flame 165 (2016) [27] Bell, D., Heyne, J. S., Won, S. H., Dryer, F. L., Haas, F. M., and Dooley, S., On the Development of General Surrogate Composition Calculations for Chemical and Physical Properties, Submitted to the 55th AIAA Aerospace Sciences Meeting, Grapevine, TX:, [28] Stouffer, S. D., Hendershott, T. H., Monfort, J., Diemer, J., Corporan, E., Wrzenski, P., and Caswell, A., Lean Blowout and Ignition Characteristics of Conventional and Surrogate Fuels Measured in a Swirl Stabilized Combustor, Submitted to the 55th AIAA Aerospace Sciences Meeting, Grapevine, TX: American Institute of Aeronautics and Astronautics, [29] Chtev, I., Rock, N., Ek, H., Smith, T., Emerson, B., Nobel, D. R., Seitzman, J., Lieuwen, T., Mayhew, E., Lee, T., Jiang, N., and Roy, S., Simultaneous High Speed (5 khz) Fuel-PLIE, OH-PLIF and Stereo PIV Imaging of Pressurized Swirl-Stabilized Flames using Liquid Fuels, Submitted to the 55th AIAA Aerospace Sciences Meeting, Grapevine, TX:, [30] Lovestead, T. M., Burger, J. L., Schneider, N., Bruno, T. J., Comprehensive Assessment of Composition and Thermochemical Variability of Three Prototype Gas Turbine Fuels by GC/QToF-MS and the Advanced Distillation-Curve Method as a Basis of Comparison for Novel Fuel Development, Energy & Fuels, in press (2016) DOI: /acs.energyfuels.6b01837 Publication Date (Web): 26 Oct [31] Outcalt, S. L., Compressed Liquid Densities of Three Reference Turbine Fuels, Energy & Fuels, in press (2016) DOI: /acs.energyfuels.6b01820 Publication Date (Web): 19 Sep [32] Edwards, J. T., Hutzler, S., Morris, R. E., Muzzell, P. A., Tri-Service Jet Fuel Characterization for DOD Applications; Task 1 Compositional Analysis/Task 2-3 Fit-For-Purpose and Trace Impurity Evaluations, May SwRI Project No [33] Yu, J., Eser, S., Determination of Critical Properties of Some Jet Fuels. Ind. Eng. Chem. Res. Vol. 34, p. 404, [34] Moses, C., A Review of ASTM D4054 Fit-For-Purpose Results, Presentation to D02.J0.06 Emerging Fuels ASTM Aviation Fuels Subcommittee, Ft. Lauderdale, FL, June 24, 2015 [35] Mueller, C. J., et al, Methodology for Formulating Diesel Surrogate Fuels with Accurate Compositional, Ignition-Quality, and Volatility Characteristics, Energy & Fuels, Vol. 26, pp , [36] Riazi, M. R. and Daubert, T. E., Analytical Correlations Interconvert Distillation Curve Types, Oil & Gas Journal, Vol. 84, 1986, August 25, pp [37] Lenoir, J. M. and Hipkin, H. G., Measured Enthalpies of Eight Hydrocarbon Fractions, Journal of Chemical and Engineering Data, Vol. 18, No. 2, 1973, pp [38] Edwards, T., Kerosene Fuels for Aerospace Propulsion Composition and Properties, AIAA Paper , July [39] Moses, C., Stavinoha, L., Roets, P., Qualification of Sasol Semi-Synthetic Jet A-1 as Commercial Jet Fuel, SwRI-8531, Nov [40] Moses, C, A Review of ASTM D4054 Fit-For-Purpose Results, Presentation to D02.J0.06 Emerging Fuels, ASTM Aviation Fuels Subcommittee, Ft. Lauderdale, FL, June 24,

44 Appendix A Tabular Property Data for Category A Fuels Table A-1 GCxGC composition (UDRI) Table A-2 Miscellaneous compositional measurements (AFPET/SwRI/UDRI) Table A-3 Physical/specification properties (SwRI/UDRI) Table A-4 Heat Capacity as a function of temperature (SwRI) Table A-5 Density, speed of sound, bulk modulus as a function of P for POSF (SwRI) Table A-6 Density, speed of sound, bulk modulus as a function of P for POSF (SwRI) Table A-7 Density, speed of sound, bulk modulus as a function of P for POSF (SwRI) 44

45 Table A-1 GCxGC composition (UDRI) GCxGC Summary Hydrogen content (weight %) Average Molecular Wt (g/mole) POSF JP-8 POSF JP-5 POSF Jet A Weight % Volume Weight % Volume Weight % Volume % % % Aromatics Alkylbenzenes benzene (C06) <0.01 < toluene (C07) C2-benzene (C08) C3-benzene (C09) C4-benzene (C10) C5-benzene (C11) C6-benzene (C12) C7-benzene (C13) C8-benzene (C14) C9-benzene (C15) C10+-benzene (C16+) <0.01 < Total Alkylbenzenes Diaromatics (Naphthalenes, Biphenyls, etc.) diaromatic-c diaromatic-c diaromatic-c diaromatic-c diaromatic-c Total Alkylnaphthalenes Cycloaromatics (Indans, Tetralins,etc.) cycloaromatic-c cycloaromatic-c cycloaromatic-c cycloaromatic-c cycloaromatic-c cycloaromatic-c cycloaromatics-c Total Cycloaromatics Total Aromatics Paraffins iso-paraffins C07 & lower -isoparaffins C08-isoparaffins C09-isoparaffins C10-isoparaffins C11-isoparaffins C12-isoparaffins C13-isoparaffins C14-isoparaffins C15-isoparaffins C16-isoparaffins C17-isoparaffins C18-isoparaffins <0.01 < C19-isoparaffins <0.01 <0.01 <0.01 < C20-isoparaffins <0.01 <0.01 <0.01 < C21-isoparaffins <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 C22-isoparaffins <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 C23-isoparaffins <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 C24-isoparaffins <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Total iso-paraffins n-paraffins n-c07 & lower n-c n-c n-c n-c n-c n-c n-c n-c n-c n-c n-c18 <0.01 <0.01 <0.01 < n-c19 <0.01 <0.01 <0.01 < n-c20 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 n-c21 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 n-c22 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 n-c23 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Total n-paraffins Cycloparaffins Monocycloparaffins C07 & lower monocycloparaffins C08-monocyclocycloparaffins C09-monocyclocycloparaffins C10-monocyclocycloparaffins C11-monocyclocycloparaffins C12-monocyclocycloparaffins C13-monocyclocycloparaffins C14-monocyclocycloparaffins C15-monocyclocycloparaffins C16-monocyclocycloparaffins C17-monocyclocycloparaffins C18-monocyclocycloparaffins <0.01 <0.01 <0.01 < C19+-monocyclocycloparaffins <0.01 <0.01 <0.01 < Total Monocycloparaffins Dicycloparaffins C08-dicycloparaffins C09-dicycloparaffins C10-dicycloparaffins C11-dicycloparaffins C12-dicycloparaffins C13-dicycloparaffins C14-dicycloparaffins C15-dicycloparaffins C16-dicycloparaffins <0.01 < C17+-dicycloparaffins <0.01 <0.01 <0.01 < Total Dicycloparaffins Tricycloparaffins C10-tricycloparaffins <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 C11-tricycloparaffins C12-tricycloparaffins <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Total Tricycloparaffins Total Cycloparaffins Average Molecular Formula - C Average Molecular Formula - H

46 Table A-2 Miscellaneous compositional measurements (AFPET/SwRI/UDRI) Property A-1 (POSF 10264) A-2 (POSF 10325) A-3 (POSF 10289) ASTM D Aromatics, %vol (AFPET) ASTM D Aromatics, %vol (SwRI) ASTM D5186 Aromatics, %mass (SwRI) ASTM D6379 Aromatics, %mass (SwRI) ASTM D6379 (UDRI) ASTM D Hydrogen Content by NMR, % mass (AFPET) ASTM D3701 H content by NMR, % mass SwRI ,

47 Table A-3 Physical/specification properties (SwRI/UDRI/AFPET) Property A-1 (POSF 10264) A-2 (POSF 10325) A-3 (POSF 10289) Density, g/l 15 C (AFPET) C (SwRI) C C C C C C C C Viscosity, cst -40 C (AFPET) C (AFPET) C (SwRI) C (SwRI) Heat of combustion, MJ/kg (SwRI) D Heat of combustion, MJ/kg (AFPET) Surface tension, dyne/cm (SwRI) D1331A -10 C C C Cetane number, ASTM D Ignition Delay (ms), ASTM D Derived cetane number, ASTM D ASTM D86 Distillation (SwRI) IBP % % FBP Flash point Freeze pt, C Smoke pt, mm

48 Table A-4 Heat Capacity as a function of temperature (SwRI) POSF POSF POSF Table A-5 Density, speed of sound, bulk modulus as a function of P for POSF (SwRI) 48

49 Table A-6 Density, speed of sound, bulk modulus as a function of P for POSF (SwRI) Table A-7 Density, speed of sound, bulk modulus as a function of P for POSF (SwRI) 49

50 Appendix B tabular property data for Category C fuels 50

51 51