Hydrocarbon absorption coefficients at the3.39-�m He–Ne laser transition

James A. Drallmeier

In view of the application of light-extinction techniques for fuel-specie-concentration measurements incombustion systems, the vapor-absorption coefficient of several hydrocarbon species at the 3.39-�mHe–Ne laser transition has been measured. The hydrocarbon species include paraffins, olefins, andaromatics. Results are included for total pressure over the range of approximately 200–650 Torr at 295K with air as the buffer gas. Observations are made regarding the difference in the absorption coeffi-cient within and between hydrocarbon classifications. © 2003 Optical Society of America

OCIS codes: 120.1740, 120.4530, 260.3060, 300.1030.

1. Introduction

Nonintrusive optical methods have enjoyed a longand ever-expanding history in the field of combus-tion. While first used primarily as diagnostic meth-ods in the combustion environment, absorption- andemission-based optical methods now are also beingimplemented as control sensors with closed-loop con-trol systems for combustion system optimization.1

Absorption-based optical techniques in the infra-red have been particularly effective in the character-ization of both gaseous combustion intermediatesand products2–4 as well as input equivalence ratiobased on fuel specie concentrations.5,6 Techniqueshave also been developed for spray combustion sys-tems that are able to discern vapor-phase fuel con-centrations in the presence of the liquid phasedrops.7–10 Critical to most of these techniques isknowledge of gas-phase absorption coefficients forquantitative results.

Many of the absorption-based techniques con-cerned with characterization of hydrocarbon fuel andcombustion intermediate species are based on an ab-sorption band near 3.39 �m for two reasons. First,since the absorption comes from C–H stretching,most hydrocarbon species have a strong band absorp-tion at this wavelength, which results in detection at

The author �drallmei@umr.edu� is with the Department of Me-chanical and Aerospace Engineering at the University of Missouri-Rolla, 1870 Miner Circle, Rolla, Missouri 65409-0050.

Received 25 June 2002; revised manuscript received 15 Novem-ber 2002.

0003-6935�03�060979-04$15.00�0© 2003 Optical Society of America

low concentrations. In addition, a coherent source ofradiation at 3.39 �m is easily achieved with a He–Nelaser, making the light source relatively inexpensive.A limitation to simple absorption techniques at 3.39�m is in distinguishing various species in a mixture,limiting the technique to measurements with single-component fuels. However, one approach to thislimitation is for one to perform a series of tests usinga range of different single-component fuels to inves-tigate the behavior of the various boiling point con-stituents in the combustion system. We have usedthis approach to study the mixture preparation inspark ignition engine fuel systems.

In developing absorption techniques principally at3.39 �m for measuring fuel-vapor concentration intwo-phase flows,10,11 we have found a lack of availablequantitative absorption-coefficient data for hydrocar-bon species. Qualitative information on the infraredspectral characteristics is available but tends to belimited to absorption-band spectral location for com-pound identification �e.g., Ref. 12�. To fill this infor-mation gap, we developed a simple apparatus tomeasure the absorption coefficient for gaseous spe-cies at the 3.39-�m transition of a He–Ne laser andpresent here some of the data developed for severalhydrocarbon species. For this work, the absorptioncoefficient was based on Beer–Lambert’s law,

IIo

� exp ��kPv L�,

where I is the intensity, Pv is the vapor concentration,and L is the path length. The absorption coefficientk is given in units of cm�1atm�1 atmospheres.

20 February 2003 � Vol. 42, No. 6 � APPLIED OPTICS 979

2. Apparatus and Test Matrix

The experimental setup is shown schematically inFig. 1. An absorption cell was constructed to holdthe gas sample while providing access for transmis-sion measurements at the desired wavelength. Thecell has a variable path length as well as being able tohandle samples from 0.1 Torr to several atmospheres.The fuel and air were premixed to known proportionsin a 1000-ml prechamber before introduction into theevacuated absorption cell. The fuel concentration inthe prechamber was determined with partial pres-sures measured by a capacitance manometer trans-ducer with an accuracy greater than 0.15% of thereading and repeatability of 0.01 Torr. The finaltotal pressure in the absorption cell was measuredwith a pressure transducer with an accuracy of �6.5Torr. Sufficient time was allowed once the gas was

introduced into the absorption cell to allow the sys-tem to reach thermal equilibrium. All measure-ments were made at 295 K.

The major components of the optical system in-cluded a 5-mW He–Ne laser operating at 3.39 �m, abeam chopper, a room temperature indium arsenide�InAs� detector, a preamplifer, and a lock-in ampli-fier. The transmission of the incident radiationwithout absorption was determined with the cellpurged with air and reduced to a pressure of less than0.1 Torr. The vapor-absorption coefficient was de-termined based on the transmission of the incidentradiation through the sample in the absorption celland the known concentration of the fuel vapor in thecell. Laser output drift during the tests was moni-tored and determined to be 1% or less. The accuracyof extinction methods is typically limited at very highor low transmission values, owing to the presence ofnoise. For this reason, transmission values werekept between 0.05 and 0.90, although accurate trans-mission values could have been obtained below 0.05,owing to the use of the lock-in amplifier.

At least three data sets were collected for each fuelspecies at different concentration levels. The rangeof concentrations used for each fuel is shown in Table1. Repeatability between tests on the resultant ab-sorption coefficient was typically better than 5%.

Since the absorption coefficient is a function of totalpressure due to pressure broadening, it was mea-sured for total pressures ranging from 200 Torr tonear atmospheric pressure. As the original applica-tion of these results was for measurements in theintake port of a spark ignition engine, this pressurerange was chosen to cover the range of pressurestypically encountered.

Validation of the system and experimental tech-nique was performed with methane. The measuredabsorption coefficients were within 5% of the valuesgiven by Edwards and Burch13 over the range of 150–700 Torr.

The nine fuels that were used in the investigationare shown in Table 1. These hydrocarbons were cho-sen because they represent single species that areoften used as engine fuels or are combined in blendsof several single components to form fuels with evap-oration characteristics that are representative ofmulticomponent engine fuels.14 The first six fuels

Fig. 1. Schematic of apparatus for determining the vapor-absorption coefficient.

Table 1. Test Fuels

Fuel Species FormulaMolecular

WeightBoiling

Point �° C�HydrocarbonClassification

Fuel Mole FractionRange �%�

2-methylbutane C5H12 72.2 27.84 Branched paraffin 0.10–0.59Hexane C6H14 86.2 68.73 Straight paraffin 0.43–0.51Heptane C7H16 100.2 98.43 Straight paraffin 0.32–0.40Iso-octane C8H18 114.2 99.24 Branched paraffin 0.34–0.73Nonane C9H20 128.2 150.8 Straight paraffin 0.15–0.17Decane C10H22 142.3 174.2 Straight paraffin 0.10–0.131-Hexene C6H12 84.2 63.48 Olefin 0.54–0.87Toluene C7H8 92.1 110.6 Aromatic 0.85–1.62o-Xylene C8H10 106.2 144.4 Aromatic 0.29–0.47

980 APPLIED OPTICS � Vol. 42, No. 6 � 20 February 2003

are paraffins �straight and branched� having single-bonded carbon structures. Hexene, an olefin, has asingle double bond between the first and second car-bons in the chain. Toluene and xylene are both ar-omatics having a six-carbon ring structure with threedouble bonds. Toluene has one additional methylgroup, while xylene has two.

3. Results and Comparison to Existing Data

The measured absorption coefficients are given inFig. 2. Most fuels show a dependence of the absorp-tion coefficient on total pressure, although for2-methylbutane, 1-hexene, and toluene the variationis within the experimental error. The variation isnot as large as has been seen in methane, probablydue to the complexity of the molecular structure ofthese larger molecules and the associated overlap ofthe spectral absorption bands.

Clear differences are seen in the absorption coeffi-cient between different hydrocarbon classifications.There is an apparent decrease in absorption coeffi-cient with increased chain length for the straight-chained paraffins �open symbols�. Although onlytwo are considered, branched paraffins �solid sym-bols� show an increase in the absorption coefficientwith increased size. There is a sizable dependenceof the absorption coefficient on pressure for the larg-est nonbranched paraffins considered, particularlydecane.

The absorption coefficient was higher for paraffinsthan for olefins and aromatics. Although only oneolefin was considered, 1-hexene had a lower absorp-tion coefficient than similarly sized �i.e., molecularweight� paraffin molecules. The absorption coeffi-cient for aromatics increased with increased molecu-lar weight but was significantly less than similarlysized paraffins. The observation of similar absorp-tion coefficients for different hydrocarbon classes overa range of molecular size and boiling point �e.g.,2-methylbutane, nonane and 1-hexene� leads to thepossibility of the construction of a blended fuel with abroader distillation curve than for a single-component fuel but with each component having asimilar absorption coefficient for quantitative fuel-airmixture engine diagnostics.

As discussed, very limited vapor-absorption coeffi-cient data are available in the open literature forhydrocarbons at this wavelength. Tsuboi et al.15

measured the absorption coefficient at 3.39 �m for afew hydrocarbons in a shock tube with the intent oflooking at temperature dependence. The broaden-ing gas used was argon, although the researchersobserved that their results showed little dependenceon the broadening gas chosen. The results for hex-ane, heptane, iso-octane and nonane �the only speciesconsidered in both studies� given here at 650 Torr,295 K was generally within the scatter of the resultspresented by Tsuboi et al. at 1 atm �taken to be 760Torr� and 298 K. Table 2 shows a comparison be-tween the results. Note that Tsuboi et al. indicateda typical error of 20% on their results but suggestedthat errors may be larger at high and low extinctions.Interestingly, their results for benzene �an aromaticspecies� showed a much weaker absorption ��0.3cm�1 atm�1� than the paraffin species, similar to theresults shown in this study for toluene and o-xylene.In addition, the absorption-coefficient results ofTsuboi et al. showed a very weak temperature depen-dence �over the range of 298–1100 K� for the hydro-carbon species in common between the two studies�hexane, heptane, iso-octane and n-nonane�. Thisimplies that for at least these fuel species, the resultspresented here could be applied over a limited tem-perature range within the level of uncertainty indi-cated by Ref. 15

Table 2 also shows the results given by Jaynes andBeam16 for common fuel species �hexane, heptane,and decane�. As shown in the table, their results areonly at low total pressures, before pressure broaden-ing takes effect. However, the values seem to be ingeneral agreement given the trends indicated in Fig.2 toward lower total pressures.

4. Summary

The determination of hydrocarbon vapor concentra-tions with optical-extinction methods in two-phaseflows and combustion systems requires knowledge ofthe vapor-absorption coefficient. This property,

Table 2. Comparison of Hydrocarbon Absorption Coefficients at RoomTemperature

FuelSpecies

Approximate TotalPressure �Torr�

k�cm�1atm�1� Reference

Hexane 650 15.4 This WorkHexane �760 15 15Hexane �11.4 18 16Heptane 650 15.1 This WorkHeptane �760 16.9 15Heptane �7.6 20 16Iso-octane 650 16.3 This WorkIso-octane �760 18.8 15n-Nonane 650 11.0 This Workn-Nonane �760 16 15n-Decane 650 6.5 This Workn-Decane �3.0 10–13 16

Fig. 2. Vapor-absorption coefficient at the He–Ne laser transitionof 3.39-�m at 295 K with air as a buffer gas.

20 February 2003 � Vol. 42, No. 6 � APPLIED OPTICS 981

which varies with the wavelength of the incident ra-diation, has been available only for a limited numberof hydrocarbons at the wavelength used by manyextinction techniques, 3.39 �m. This work providesthe absorption coefficient for several hydrocarbons atthe He–Ne laser transition of 3.39 �m. The resultsare presented for single hydrocarbon species, includ-ing paraffins, olefins, and aromatics, over a range oftotal pressures typically encountered in the intake ofspark ignition engines �the application at hand�.The absorption coefficient was measured at ambienttemperature; however, previous work has shown lit-tle dependence of the absorption coefficient at thiswavelength for many hydrocarbons over reasonabletemperature ranges. Observations are made con-cerning the correlation of absorption coefficient tohydrocarbon classification.

References1. N. Docquier and S. Candel, “Combustion control and sensors:

a review,” Prog. Energy Combust. Sci. 28, 107–150 �2002�.2. D. S. Baer, V. Nagali, E. R. Furlong, and R. K. Hansen,

“Scanned- and fixed-wavelength absorption diagnostics forcombustion using multiplexed diode lasers,” AIAA J. 34, 489–493 �1996�.

3. M. G. Allen, “Diode laser absorption sensors for gas-dynamicand combustion flows,” Meas. Sci. Technol. 9, 545–562 �1998�.

4. B. Mizaikoff, P. Fuss, and M. J. Hall, “Fast-spec: an infraredspectroscopic diagnostic to measure time-resolved exhaust hy-drocarbon emissions from S.I. engines,” in Proceedings of theTwenty-Seventh Symposium �International� on Combustion,2093–2100 �The Combustion Institute, Pittsburgh, Pa., 1998�.

5. M. Golombok, P. J. Cooney, J. S. Kitching, and S. R. Nattrass,“Spatial resolution of fuel distribution in an engine using in-frared absorption,” Rev. Sci. Instrum. 68, 4236–4246 �1997�.

6. J. G. Lee, K. Kim, and D. Santavicca, “Measurement of equiv-alence ratio fluctuation and its effect on heat release duringunstable combustion,” in Proceedings of the Twenty-EighthSymposium �International� on Combustion, 739–746 �TheCombustion Institute, Pittsburgh, Pa., 2000�.

7. A. R. Chraplyvy, “Nonintrusive measurements of vapor con-centrations inside sprays,” Appl. Opt. 20, 2620–2624 �1981�.

8. E. Winklhofer, G. K. Fraidl, and A. Plimon, “Monitoring ofgasoline fuel distribution in a research engine,” Proc. Inst.Mech. Eng. D 206, 107–115 �1992�.

9. J. A. Drallmeier, “Hydrocarbon vapor measurements in fuelsprays: a simplification of the infrared extinction technique,”Appl. Opt. 33, 7175–7179 �1994�.

10. P. D. Jennings and J. A. Drallmeier, “Pulsed fuel spray vaporphase characterization,” Atomization Sprays 6, 537–562�1996�.

11. J. H. Wright and J. A. Drallmeier, “Cyclic variability of pulsedspray vapor fields,” Exp. Fluids 25, 329–336 �1998�.

12. American Petroleum Institute, American Petroleum InstituteResearch Project 44 Infrared Spectral Data, 2,2,4 Trimeth-ylpentane, Contributed by the U.S. Naval Research Labora-tory, Washington, D.C. �1947�.

13. B. N. Edwards and D. E. Burch, “Absorption of 3.39-micronhelium-neon laser emission by methane in the atmosphere,” J.Opt. Soc. Am. 55, 174–177 �1965�.

14. J. P. Styron, P. L. Kelly-Zion, C. F. Lee, J. E. Peters, and R. A.White, “Multicomponent liquid and vapor fuel distributionmeasurements in the cylinder of a port-injected, spark-ignitionEngine,” SAE Technical Paper Series, Paper 2000-01-0243 �So-ciety of Automotive Engineering, Warrendale, Pa., 2001�.

15. T. Tsuboi, K. Inomata, Y. Tsunoda, A. Isobe, and K. Nagaya,“Light absorption by hydrocarbon molecules at 3.392 micronsof He–Ne Laser,” Jpn. J. Appl. Phys. 24, 8–13 �1985�.

16. D. N. Jaynes and B. H. Beam, “Hydrocarbon gas absorption bya HeNe laser beam at a 3.39 � wavelength,” Appl. Opt. 8,1741–1742 �1969�.

982 APPLIED OPTICS � Vol. 42, No. 6 � 20 February 2003