prepared for - nasa · 2013-08-31 · h 7 3 2 3 0 9 8 nasa cr-121159 adl 74443 odor intensity and...

H 7 3 2 3 0 9 8NASA CR-121159

ADL 74443

ODOR INTENSITY AND CHARACTERIZATION OF

JET EXHAUST AND CHEMICAL ANALYTICAL MEASUREMENTS

DAVID A. KENDALL and PHILIP L LEVINS, Ph.D.

ARTHUR D. LITTLE, INC

prepared for

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

NASA Lewis Research Center

Contract NAS 3-15701

Arthur D Little, Inc.

https://ntrs.nasa.gov/search.jsp?R=19730014366 2020-05-20T17:53:17+00:00Z

1. Report No. 2. Government Accession No.

NASA CR-1211594. Title and Subtitle ODOR INTENSITY AND CHARACTERIZATION OF JET

EXHAUST AND CHEMICAL ANALYTICAL MEASUREMENTS.

7. Author(s)

David A. Kendall and Philip L. Levins, Ph.D.

9. Performing Organization Name and Address

Arthur D. Little, Inc.Acorn ParkCambridge, Massachusetts 02140

12. Sponsoring Agency Name and AddressNational Aeronautics and Space AdministrationWashington, D. C. 20546

3. Recipient's Catalog No.

5. Report DateMarch 1973

6. Performing Organization Code74443

8, Performing Organization Report No.

ADL- 74443

10. Work Unit No.

N/A

11. Contract or Grant No.

NAS 3-1570113. Type of Report and Period Covered

Contractor ReportMarch 1972-March 1973

14. Sponsoring Agency Code

15. Supplementary Notes

Project Manager, Helmut F. Butze, Air Breathing Engines Division, NASA-LewisResearch Center, Cleveland, Ohio .

16. Abstract

Odor and chemical analyses were carried out on the exhaust samples from a J-57combustor can operated over a range of inlet conditions, and with several fueltypes and nozzle modifications. The odor characteristics and total intensityof odor for each exhaust were determined by a trained panel of NASA personnelover a range of dilutions to allow for a least-squares determination of theintensity at 1,000 to 1 dilutions. Analytical measures included the concentrationof total hydrocarbons (by on-line flame ionization detector) and the concentrationsof aromatic organic species and oxygenated organic species from collected sampleswhich were taken concurrently. A correlation was found between the concentrationof the odorous oxygenated fraction and the total intensity of aroma. Inlet operatingconditions and nozzle modifications which increase the efficiency of combustion asmeasured by exhaust gas analyses reduce the odor intensity and the quantity ofoxygenates in the exhaust. The type of fuel burned altered the intensity of odor inrelation to the quantity of oxygenates produced and, in some instances, changed theodor character. A description of the equipment and procedures used is included.

17. Key Words (Suggested by Author(s))

Combustor Exhaust Odor

Odor Analysis

18. Distribution Statement

Unclassified - unlimited

19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price*

Unclassified Unclassified 58

: For sale by the National Technical Information Service, Springfield, Virginia 22151

NASA-C-168 (Rev. 6-71)

FOREWORD

The authors wish to recognize the contribution of the NASA personnelwho voluntarily served on the Odor Panel:

Evelyn AnagnostbuRobert BranstetterAndrew LeavittRobert ManlyPatricia O'DonnellRobert SummersJohn TomaErline TrsekH. F. Butze, Coordinator

Their participation and cooperation was essential to the success of theprogram.

The Arthur D. Little, Inc., Staff members and their areas of respon-sibility are as follows:

Odor Studies

Shirley RaymondRalph" SteevesFredrick Sullivan

Analytical Studies

Alegria CaragayRafael Cruz-AlvarezJames Stauffer

Mathematical Analysis

Janet CappersDr. Irwin MillerCarol Ulmer

iii

Arthur D Little, Inc

TABLE OF CONTENTS

Page No.

I. SUMMARY ' : 1

II. INTRODUCTION 3

Objective 3

Experimental Approach 4

III. EXPERIMENTAL PROGRAM 6

Odor Test Facility - Design and Construction 7

Panel Selection and Odor Measurement 15

Exhaust Sampling and Chemical Analyses 18

Combustor Test Program 24

IV. RESULTS 27

Preliminary Testing 2?

Odor Survey of the Effects of Combustor Operating Parameters 28

Effects of Combustor Modification and Operating Variables 32

Fuel Variables 38

Odor Character and Composition Variation . 41

V. DISCUSSION OF RESULTS 45

Odor Analytical Correlation 45

Combustor Variables 46

Fuel Variables 49

Efficiency and Emission Levels 49

Odor Character and Composition Variation 53

VI. CONCLUSIONS 56

Recommendations 56


LIST OF TABLES

Table No. Title Page No.

1 EXHAUST CONCENTRATIONS FOR ODOR TESTS 14

2 EXHAUST CONCENTRATIONS FOR HIGH DOSE ODOR TESTS 14

3 PREPARATIVE COLUMN CHROMATOGRAPHY OF EXHAUST 23SAMPLES

4 HIGH RESOLUTION MASS SPECTROMETRIC ANALYSIS OF 25

OXYGENATE FRACTIONS (LCO) (where ASTM-A1 fuelwas burned)

5 OPERATING CONDITIONS AND EXHAUST ANALYSES 27

6 OPERATING CONDITIONS AND EXHAUST ANALYSES 29ASTM-A1 FUEL/DUPLEX NOZZLE



9 OPERATING CONDITIONS AND EXHAUST ANALYSES 36ASTM-A1 FUEL/AIR ASSIST NOZZLE

10 OPERATING CONDITIONS AND EXHAUST ANALYSES 37ASTM-A1 FUEL/SIMPLEX NOZZLE

11 OPERATING CONDITIONS AND EXHAUST ANALYSES . 39NATURAL GAS FUEL/SPECIAL NOZZLES

12 OPERATING CONDITIONS AND EXHAUST ANALYSES 40ISOOCTANE FUEL /DUPLEX NOZZLE

13 OPERATING CONDITIONS AND EXHAUST ANALYSES 42JP-4 FUEL/DUPLEX NOZZLE

14 OPERATING CONDITIONS AND EXHAUST ANALYSES 43JP-4 TYPE FUEL/DUPLEX NOZZLE

15 ODOR PROFILE COMPOSITES 44EXHAUST CHARACTER ASSOCIATED WITH FUEL TYPE

16 TIA INTERCEPT AND OBSERVED LCO VALUES FOR 47

VARIOUS MODIFICATIONS AT VARIOUS INLET AIRTEMPERATURES

vi


LIST OF TABLES (CONT'D.)

Table No. Title Page No.

17 TIA INTERCEPT AND OBSERVED LCO VALUES FOR 50VARIOUS FUELS AT FOUR INLET AIR CONDITIONS

18 EMISSION LEVELS OF ODORANT SPECIES AST^Al 54

vii


LIST OF FIGURES

Figure No. Title Page No.

1 SCHEMATIC OF THE ANALYTICAL PROCESS 5

2 SCHEMATIC ARRANGEMENT OF EQUIPMENT FOR NASA ODOR 8TEST FACILITY

3 CUTAWAY VIEW OF NASA ODOR TEST FACILITY 9

4 NASA PANEL INSIDE ODOR TEST FACILITY H

5 EXHAUST SAMPLING AND DILUTION SYSTEM 12

6 REAR VIEW OF ODOR TEST FACILITY 13

7 ODORIZATION SYSTEM FOR ODORANT SOLUTIONS 16

8 EXHAUST SAMPLING TRAPS 19

9 ALC RESEARCH PROTOTYPE SYSTEM 21

10 ALC/UV CHROMATOGRAM OF JET EXHAUST TOE SAMPLE 22

11 TIA vs yg/LCO FOR ODOR SURVEY - ASTM-A1 FUEL/ 31DUPLEX NOZZLE

12 TIA vs yg/LCO FOR ALL NOZZLE VARIATIONS 48

13 TIA vs yg/LCO FOR ALL FUELS 51

14 TIA vs INEFFICIENCY FOR TESTS WITH ASTM-A1 FUEL 52

viii


SYMBOLS

The following terms have been used frequently in the text and aredefined for the convenience of the reader.

Coordinating Research Council

Environmental Protection Agency

National Aeronautics and Space Administration

Inlet air temperature, °C

Inlet air pressure, atm.

Air reference velocity, m/sec

Fuel/air ratio

Total intensity of aroma. This is recorded on a6-point scale in % units with 3 being a maximum or strongodor level. Frequently tabulated as the value computedfrom the dose/response data at a 1000-to-l dilution(1 JZ./m3) for exhaust or odorous fraction.

HRMS - High-resolution mass spectrometry.

R&DB

CRC -

EPA -

NASA -

Ti -P.i

V

f/a -

TIA -

UV

Rings and double bonds - a measure of the hydrogenunsaturation determined from the elemental composition' by HRMS.

Ultraviolet absorption. The detector system used inthe ALC recording optical density at 254 nm.

FID - Flame ionization detector for measuring hydrocarbons.

ALC - Analytical liquid chromatography - the basis for theinstrumental method under development.

NA - Preparative scale sampling on Chromosorb 102500SL of exhaust sampled.

AN - Analytical scale sampling on Chromosorb 102 - generally

30A of exhaust sampled.

THC - Total hydrocarbons as measured by FID.

TOE - Total organic extract - the total organic exhaustextract of components recovered from the sample collectionby solvent extraction.

LCP - The paraffin fraction isolated from the preparativeliquid chromatography procedure.

IX


SYMBOLS (Cont'd.)

LCA - The aromatic (oily kerosene) fraction isolated fromthe preparative liquid chromatography procedure.Concentration of aromatics measured by ALC.

LCD - The oxygenated (smoky-burnt) fraction isolated fromthe preparative liquid chroma tography procedure.Concentration of oxygenates measured by ALC.

pg/Ji - Concentration of exhaust species reported as yg/&of exhaust; mg/m^ or mg/k£ are equal to pg/£.

r2 - A statistical measure of the % of information explainedby the dependent variable.

a - Standard error; a measure of the variability of the data.


I. SUMMARY

In March 1972, an experimental program was undertaken by Arthur D. Little,Inc., for Lewis Research Center, The National Aeronautics and Space Admin-istration, to define the odor character and intensity of jet combustorexhaust, and to investigate the chemical relationships between odor in-tensity and odor character. It was planned to study and compare thesensory characteristics of exhaust from a combustor operated at a varietyof inlet air conditions, with several different fuels and two or "morenozzle modifications. The odor measurements were made by a panel ofNASA personnel who had been trained by ADL to describe the odor character-istics and measure the odor strength. An odor test facility was designedand constructed to present exhaust samples over a range of concentrations.This technique for odor measurement summarizes the odor intensity of agiven sample of exhaust by determining the least-squares line for therange of intensities and concentrations expressed as the logarithm. Fromthis line, the total intensity of aroma for one liter of exhaust per cubicmeter of dilution air can be calculated. This is the odor strength ofeach exhaust at a 1000-to-l dilution.

Since preliminary experiments early in the program indicated that thetechniques developed during our studies of diesel exhaust were applicablefor sampling and analysis of jet combustor exhaust, little effort wasexpended in carrying out detailed identification studies of the verycomplex mixture of components which is the organic phase of jet combustorexhaust. However, analytical measurements based on the diesel exhaustwork were taken at each experimental condition. These included the col-lection of the organic components from the exhaust by adsorption and theirextraction from the adsorbent with solvent. The total organic extractwas separated and quantified using analytical liquid chromatography todetermine the aromatic components, representing the fuel> and the oxygen-ated species, representing the more odorous partial combustion products.In addition, a more detailed analysis of the relative distribution ofvarious chemical species, based on their chemical compositions, was con-ducted by means of high-resolution mass spectrometry. This analysis wasundertaken to assure that the chemistry of the odorants was consistentwith that previously determined for diesel exhaust. In addition, itwas felt that such information might provide a means for differentiatingbetween exhausts of significantly different odor types.

Finally, Standard exhaust emission measurements were taken during thetest run with on-line monitoring equipment operated by NASA personnel.These measurements monitored the inlet and exit operating parameters,including the concentrations of total hydrocarbons, CO, CC>2, NO, andNOX as well as the exhaust temperature profile. All of this informa-tion was automatically recorded in the NASA data acquisition system.From these data the thermal efficiency of combustion (eta) and theefficiency based on exhaust composition (CEFF) were calculated.


As a result of these studies, a significant correlation was found betweenthe total intensity of aroma for combustor exhaust and the logarithm ofthe partially oxygenated components, which extends over all operatingconditions, nozzle modifications, and fuel types. Similar semi-logarithmiccorrelations were found with the concentration of the aromatic componentsand, to a lesser degree, with the total hydrocarbons as measured by theon-line system. In addition, an inverse correlation was found betweenthe logarithm of the efficiency of combustor operation, as measured byexhaust gas analyses, and the intensity of the exhaust odor. At high-efficiency operation, the odor level is significantly lower than at low-efficiency operation. Minimum odor levels were observed when the effic-iency fell in the 99.8 to 99.9% range. Combustor modifications whichimproved the efficiency of combustion (such as the air assist nozzle atlow fuel flows and inlet operating conditions, or the Simplex nozzle athigh fuel flows and inlet air conditions) significantly reduced the odorintensity as well as the concentration of oxygenates in the exhaust.

Different fuels were also found to affect the intensity of odor. Thisvariation in intensity is proportional to the oxygenates produced.Therefore, although the intensity of odor at all operating conditionsproved to be substantially lower with natural gas as the fuel, theconcentrations of oxygenates measured were proportionately low as well.A reduction of odor was also observed when isooctane was used as a fuel,again in relationship to the quantity of oxygenated species produced.

Differences were noted in the character of the exhaust, particularlybetween fuels. Most noticeable were the characteristics associatedwith isooctane, which was described as "burnt sweet" rather than "smokyburnt", and was less pungent. When natural gas was burned the levelswere substantially lower in odor intensity, but also were associatedwith burnt sugar descriptors and, at one condition, with irritants suchas acrolein.

-2 -


II. INTRODUCTION

During the past twelve months, Arthur D. Little, Inc., has carried outa study of the odor characteristics and their intensities from jet corn-bus tor exhaust gases under contract to, and in conjunction with, NASA-Lewis Research Center. NASA-Lewis provided the test facility, operators,and on-line monitoring equipment of the exhaust gases. In addition, per-sonnel from NASA-Lewis were trained to carry out the odor measurementsunder the direction of the ADL staff. Analytical aspects of the programwere carried out by the ADL staff, both at NASA-Lewis and in Cambridge,Massachusetts, using methodology developed primarily as a result of ADL'swork for the Coordinating Research Council (CRC) and the EnvironmentalProtection Agency (EPA) on diesel exhaust odor.1*2* 3>lt Although con-siderable effort has been expended on analyzing the odors of diesel ex-haust during the past 10 years, including work at Southwest ResearchInstitute, Illinois Institute of Technological Research Institute, ScottLaboratories, and the Bureau of Mines, to mention a few, no detailedstudies of jet combustor or jet turbine odor levels have been conducted.

The program was originally outlined in three tasks: (1) the orientationand design phase, (2) the construction of the odor test facility andtraining of an odor panel, and (3) the examination of the effects of com-bustor variables on the intensity and odor character of the exhausts. Thisprogram resulted in the development of useful and significant correlationsbetween levels of certain organic species and the intensities of odor, aswell as between odor intensity and the efficiency of combustor operation.The success of the program was due, in large part, to the technology andexpertise developed in the course of the five years of work with CRC/EPA.

Objective

The statement of work of the original program had dictated a study programto define the chemical nature of the odorous components of jet combustor ex-haust under various operating conditions, and to measure the intensity of odorand its character. Since 'a significant amount of the analytical informationneeded had been developed in the first three years of our study of dieselexhaust, repetition of the detailed analytical work did not appear to bewarranted. The technology had been developed to a point at which certainchemical measures which correlated with diesel exhaust odor intensity hadbeen taken and defined. The study was directed at determining that thesemeasures also related to jet combustor odor levels. The objective of thesensory analysis program was to obtain measures of intensity based on avariable dose presentation of the exhaust, to relate to concurrent measuresof the concentrations of odorous components. Additionally, the panel wasasked to describe the character of the odor, particularly as it differedbetween test variables. Samples of sufficient size were taken so that

- 3 -


odorant presentations could be made at subsequent times and more detailedanalyses could be made of the kinds and relative concentrations of chem-ical species present.

The objective of the experimental program, therefore, was to (1) measure

the odor intensity of jet combustor exhaust under a wide variety of operatingconditions, (2) characterize the odor, types associated with either fuelor operating conditions, (3) measure chemical concentrations in the ex-haust to correlate with the intensity measure, and (4) probe the vari-ation in the chemical composition of the complex mixture represented inthe oxygenate fraction derived from jet combustor exhaust.

Experimental Approach

The experimental studies > aside from the tasks outlined in the WorkStatement, were carried out in three phases. An initial exploratoryrun was carried out in the first month of the program to demonstratethe applicability of the techniques developed in ADL's earlier studiesof diesel exhaust to the collection and analysis of jet combustor ex-haust. When these techniques proved to be satisfactory, a survey ofthe extremes of the engine operating parameters using ASTM-A1 fuel wascarried out during the months of July and August. From this work, itappeared that essentially two levels 'of odor could be produced by thecombustor: a higher level of odor associated with low-efficiency oper-ation,and a lower level of odor associated with high-efficiency operation,Therefore, during the third phase of the experimental work (which repre-sented the bulk of the effort) three combustor-inlet conditions wereexamined with a variety of fuels and combustor modifications. At theconclusion of the program, three additional inlet air conditions ofpotentially lower operating efficiency were examined. During the thirdphase of the program, each variable was replicated on separate days inits entirety, and for each test point, duplicate samples were taken foranalysis while odor measurements were made both in ordered and randompresentation.

The. inlet air conditions in the survey program were varied from ,10-240 Cinlet air temperature, 1-4 atm. inlet air pressure, 15-36 m/sec airreference velocity, and 0.008 to 0.016 fuel/air ratio. . .

Exhaust samples were presented to a trained panel of NASA personnel andconcurrently sampled for chemical analysis. A schematic diagram of theprocess appears in Figure 1.

_ 4 _


ChromosorbTrap

Odor PanelAnalyses

HydrocarbonsCOC02NO

PreparativeNA

(500£)

AnalyticalAN(304)

NO,, AnalyticalLiquid

GhromatographyALC/UV(all samples)

SolventExtractTOE

PreparativeScale

Liquid Chromatography

ParaffinsLCP

AromaticsLCA

OxygenatesLCO

NoOdor Mass

OdorSpectrometry

Odor.High Resolution Mass

Spectrometrv

FIGURE 1 SCHEMATIC OF THE ANALYTICAL PROCESS

- 5 -


III. EXPERIMENTAL PROGRAM

The purpose of the experimental program was to define the odor characterand intensity of jet combustor exhaust and to understand the chemistrythat contributes to its odor character and intensity. Such a programrequires detailed and reliable odor data which can be correlated withanalytical measurements of sufficient sensitivity to determine the con-centrations of those components that influence the odor sense. To alarge extent, the analytical side of the program had been developed as aresult of the several years of work performed by ADL on diesel exhaustodor under the sponsorship of CRC/EPA, and it remained to automate andimprove the analytical system to provide reliable information of theconcentrations of components in the yg/£ range. In the diesel exhaustprogram, ADL's staff of highly trained and experienced odor analystswas used to develop the analytical correlations and an understandingof the chemistry of diesel exhaust odor. In the present program, NASApersonnel were used to implement the odor measurements. This requiredtraining and indoctrinating a group of volunteers to provide the neces-sary information.

The study of odors is, at best, a difficult problem in that differentchemicals and combinations of chemicals may be perceived as differentodor characteristics or odor types. However, an individual's memoryfor odor character is good, and he can be trained to recognize anddifferentiate between thousands of different odor types. Odorant chem-icals at different concentrations produce different strengths of per-ceived odor. This is referred to as the odor intensity. It is wellrecognized that the perceived sensation is a logarithmic function ofthe concentration,5 and it is important to remember this logarithmicrelationship to odor strength in understanding the requirements of ananalytical program. In addition, odorants evoke what is frequentlytermed an hedonic response; i.e., people either like or dislike theodor type. This influences the way an individual responds to the inten-sity of odor, and thus the relative strength of an odor that is reportedmay not reflect a difference in concentration. For example, a low inten-sity of butyl mercaptan may be judged and reported to be strong becauseof the unpleasant character of the skunk odor. A similarly low level ofcitral may be judged to b$ pleasant, and therefore, a slight intensitywould then be recorded, whereas the actual odor strengths of the twocompounds could be deemed approximately the same. One function of theodor training program is to minimize the effects of such like/dislikejudgments in reporting intensities.

In measuring odor it is essential to have a presentation system whicheliminates distractions, either from noise or interruptions, as well asextraneous odors. The system must also provide for reproducibly con-trolling the concentration of the odorant dose presented.to the panel.

- 6 -


Much odor intensity information has been obtained on the basis of presen-tation of the odorant at a single concentration, using an intensityscale as the means of scoring differences. This technique has two prob-lems: first, it can be shown that at both low and high concentrationsthe discrimination between changes in concentration is less precise thanthe intensity discrimination in the middle portion of the intensity range;and, second, by obtaining a series of responses from a set of concentra-tions, the intensity measure can take into account the influence of thechange in perceived intensity with increasing concentrations. Theseries of measurements can then be summarized into a single data pointby means of a least-squares regression analysis of the dose responsecurve for the individual panel members or the panel together. This resultsin an equation of the form TIA = a + b logC, where C is exhaust concen-tration in £/m3. The intercept (a) is the TIA, where C=l Jl/m3 and (b) isthe slope of the dose/response relationship.

A small panel was trained, to make odor measurements at five test concen-trations, usually from 0.15-3.6 Vm3, with each test about double theprevious concentration.

A regression analysis has least variability in the middle portion ofthe data. It was, therefore, mathematically correct and convenient toselect a value to represent the total intensity of exhaust as the TIAcalculated for exhaust at a concentration of 1 £/m3. Although the con-centration is not precisely in the center of the dose/response data, itis close to it. Since all of the analytical measurements are recorded asyg/£ of exhaust, and since the TIA is calculated for 1 liter of exhaust •per cubic meter of dilution air, the analytical concentration can bedirectly substituted to give the concentration of odor-active componentsas yg/m3 of dilution air. (This is a dilution of about 1 ppb on avolume basis.)

The following sections of the report describe the efforts related tothe construction of an odor test facility at NASA-Lewis; panel selectionand methodology of odor measurements, the exhaust sampling and chemicalanalytical procedures for measurement of odorous species, and finally thecombustor operation and test program.

Odor Test Facility - Design and Construction

The odor test facility consists of a chamber approximately 2.4 meterswide, 3.6 meters long, and .2.4 meters high, with a plenum chamber locatedon the rear wall in which exhaust can be diluted to the desired concen-tration with activated carbon-treated air, and exhausted to the outsidewithout contaminating the air in the vicinity of the test chamber.Figure 2 presents the layout and flow patterns of the facility, whileFigure 3 shows the orientation of equipment. To provide low ambient

- 7 -


(1)

108£/sec Make-up Air

Test Chamber

Plenum

Adsorbent0.50 £/sec

Pressure Gauge

Jet Combustor

-{XH!

HXH

HotFilter

Mixer

0.19£/sec

70fi/secCooled and

Carbon-Treated Air

2.36 2/sec Sample

IAnalyzer Console(THC, CO, NO,

,C0 2 ,NO X )

Orifice

FIGURE 2 SCHEMATIC ARRANGEMENT OF EQUIPMENTFOR NASA ODOR TEST FACILITY

Arthur D Little Inc

I I .

d

HCO

OS

8

O

1H

U

- 9 - Arthur D Little Inc

odor in the test facility and for the comfort of panel members duringthe analysis, the air is cooled and carbon-treated for both the testchamber and for the plenum. The air from the test chamber is recycledwith make-up air added at the intake to the air conditioner. The testchamber is maintained at a positive pressure with respect to the plenumchamber to produce flow into the exhaust stream when the sniffing portsare open. The sniffing ports (Figure 4) located side by side along therear wall of the test chamber, are normally sealed with an aluminumplate and a deodorized rubber gasket material. The hole which has beencut through the gasket is shaped to accommodate the analyst's face sothat the nose is in the air stream while air flow around the face is min-imized. The flow of air in the plenum chamber is from bottom to top atlinear flow rate found to be approximately 15 Jl/sec. This air flow isnot uncomfortable for the analyst, and variability in flow rate does notaffect the concentration of exhaust in the air stream.

During the test day, air is continuously cycled through the test chamberand through the plenum chamber, from which it is exhausted at about108 H/sec. The recirculated air acts to maintain a low odor level inthe test chamber. The exhaust system removes the odorized sample and exr-traneous odors from the plenum chamber and part of the air from thetestroom to minimize the build-up of odor at the sniffing stations.Little build-up of odor was observed in the test chamber, and with airconditioning the facility remained quite operable, even with 30°to 35°Cambient air temperatures. Total air flow to the plenum chamber wasmeasured at 70 £/sec, using a vane-type anemometer, which was approxi-mately the diameter of the input duct.

A small constant flow of this air is taken for dilution air by means ofa 2.36 H/sec metal bellows pump. To produce a range of exhaust con-centrations at dilutions appropriate for odor analysis, a smaller, 0.19H,/sec, sampling pump takes exhaust through one of five orifices to givefive predetermined and controlled flows of exhaust. Figure 5 shows theexhaust sampling and dilution system. The diluted exhaust was mixedwith the total air flow to the plenum by means of an orifice-type mixingtube. This can be seen in the rear view photograph of the test chamberin Figure 6. The delivery rate of the 2.36 H/sec pump was found to be1.98 H/sec at atmospheric inlet pressure, which gives a constant 35-to-1 dilution for all samples.

Concentration ratios, obtained for the five orifices presently in use,are given in Table 1.

- 10 -


I

HI

CC

LUO

co

UJ

1

cc

- 11 -Arthur D Little Inc

CO

O\-

_i5a<(3

I

xLU

inLJJcc

- 12 -

Arthur D Little Inc

o<

V)01

cc§O

(O

LUCCDCD


TABLE 1

EXHAUST CONCENTRATIONS FOR ODOR TESTS

ExhaustTotal Concentration

Test Condition Dilution &/m3

TI 6600/1 0.15

T2 3300/1 0.30

T3 1600/1 0.62

T. 700/1 1.44T 280/1 3.6

The system was designed to produce five concentrations, each one approx-imately double the previous concentration. As can be observed from thedilutions calculated from CO measurements at the smaller diameters, theeffect is that of a critical orifice, and the ratio holds fairly well.At the two higher concentrations, the doubling is not exact but this istaken into consideration in the mathematical analysis of the data, al-though it introduces some bias.

To obtain higher concentrations of exhaust, the system is equipped withtwo sets of orifices to control the relative flows of exhaust and dilu-tion air to the 2.36 H/sec pump without the use of the 0.19 £/sec pump.Dilution ratios in the range of 200/1 to 35/1 can be produced by select-ing the appropriate pair of orifices for the exhaust sampling line andthe diluting air stream (Table 2).

TABLE 2

EXHAUST CONCENTRATIONS FOR HIGH DOSE ODOR TESTS

Exhaus tConcentrations

Test Condition Dilution A/m3

Tm 180/1 5.6

T2R 110/1 9.0

T.u 80/1 12.5jnT4H 71/1 14.0

T 36/1 28.0

- 14 -


To assure complete odor removal, a standard three-minute interval iswaited between all tests. Experience has indicated that it requiresapproximately 20 seconds for the mixed exhaust and dilution air sampleto reach the sniffing ports after the valve has been opened from theexhaust line to the 0.19 I/sec pump. The odor test itself requiressomething less than 30 seconds for all four panel members to make theiranalyses and record odor intensity and character.

A second sampling system has been developed to permit the injectionof diluted samples of the total organic extract (TOE) collected in thepreparative scale sampling procedure. A Harvard syringe pump (Figure 7)is located so that the syringe needles can be inserted into a heatedsegment of a 1.27 cm stainless-steel line on the outlet of the 2.36 I/secpump (or later a 0.50 A/sec pump). A uniform flow of a solution of theTOE at varying rates can be perfused into the air stream during a timedsequence. At No. 6 gear setting, the pump is found to deliver a maximumof 2 yJl/sec. At this gear setting the feed rate can be continuouslyvaried down to less than 10% (0.2 p£/sec). Thus, the concentration rangeof raw exhaust presented can be approximated by controlling the dilutionof the TOE in a low odorous solvent, such as nonane.

Panel Selection and Odor Measurement

A major portion of the program depended on the development of reliableodor information, both descriptive as to type and quantitative as tostrength. Since odor is the sensory response to certain chemicals,called odorants, it is necessary to use human subjects and train themto be objective in their measurements. Therefore, one day was spentin selecting eight interested and capable individuals from a group ofsixteen volunteers from Lewis Research Center.

Panel Orientation

The eight panelists were given three days of training which includedpractice in the identification of different odor types related to dieselexhaust and scaling the odor strength perceived with different concentra-tions of odorants. Reference standards for use in describing odor typeswere taken from ADL's library of odorants as well as mixtures of chemicalsand exhaust extracts derived from the diesel exhaust studies.

The intensity scale used by the panel is a simple six-point scaledescribed in Methodology of the Flavor Profile. It represents an individual'sability to discriminate between slight, moderate and strong odor levels,which are designated as 1, 2, and 3. With training, the panel memberscan estimate intermediate levels with reasonable reliability, and these

- 15 -


FIGURE 7 ODORIZATION SYSTEM FOR ODORANT SOLUTIONS


are designated as % integers. Our experience indicates that theconcentration difference between slight and moderate intensities may beas much as a tenfold increase in concentration. It is possible for anindividual to reproducibly measure substantially smaller differences inintensity, but only where direct comparisons can be made without sensoryfatigue.

Although the ability to rate intensity developed rather rapidly, thedevelopment of descriptive vocabulary was a continuing process throughoutthe program. The performance of the panel, even during the odor surveywhich was their first task, was good and the results were uniformly ofvalue in understanding the effect of operating variables on odor.

Odorant Presentation

Based on a number of odor studies carried out at ADL, and the primaryconcern for determining reliable intensity data, exhaust from each testwas presented in a series of five concentrations, each approximatelytwice the preceding level. Each panelist independently recorded thetotal intensity of aroma (TIA) and the primary odor character notes.This required a minimum number of sniffs (2-3) and less than 30 secondsat each concentration. After completion of one dose/response series inorder of increasing concentration, the same doses were presented inrandom order. At least three and preferably four NASA panelists par-ticipated at each session with one ADL staff member acting as panelleader.

Data Analysis

The TIA data is most easily summarized by calculating the least-squaresline for TIA vs log of the exhaust concentration in liters per cubicmeter. The TIA intercept, i.e., the TIA calculated for 1 liter ofexhaust per cubic meter of dilution air, represents the information from20 to 25 observations for both the ordered and the random sets. Thesesets were kept separate so that the effects of the two different modesof presentation could be assessed.

- 17 -


The summarization of the descriptive information requires some interpre-tation and weighting of individual's experience and vocabulary. Sincedescriptive information is recorded for all presentations, but usuallyin greater detail for the ordered series, the descriptive odor profilespresent the most frequently used reference terms and eliminate wordingthat is redundant. Of primary interest is the description of combustioncomponents "smoky", "burnt", "sooty", and "oxidized oil", as well as fuelrelated odor characteristics "solventy", "kerosene", "naphthalene","oily". In addition, the irritation factors are tabulated. The profilesummary represents the odor characteristics at an intermediary concen-tration rather than the high or low doses.

Exhaust Sampling and Chemical Analyses'

Sampling Methods

Jet combustor exhaust samples from which the particulates have been re-moved by filtration through a heated fiberglass filter are collected onpreparative traps and analytical traps. The preparative trap (Figure 8)contains 10 g of prewashed Chromosorb® 102 (60/80 mesh, Johns Manville)and is used for the collection of 500£ of exhaust. The analytical trapsare used for the collection of small amounts of exhaust, e.g., 30& .These are made from a 7.5 cm piece of 0.8 cm O.D. stainless-steel tubingfitted at each end with 0.4 x 0.8 cm Swagelok® reducing union and holdabout 1 g of Chromosorb® 102.

The organic components of exhaust adsorbed by the Chromosorb® are sub-sequently extracted by allowing freshly distilled chromatographic-gradepentane to slowly percolate into the Chromosorb® bed at the rate ofapproximately 0.5 m£/min. The solvent flow is countercurrent to thedirection of exhaust flow to the trap during sample collection. Theextracts so derived are known as the total organic extract (TOE). TheTOE fractions contain essentially all of the odor. Ten ml of pentaneeffluent are required for the complete extraction of the exhaust samplesfrom the preparative traps, while 1 mA of effluent is sufficient for theanalytical traps. Therefore, the extract from the preparative trapis equivalent to 0.05£ of exhaust/p£ of solution, while the extractfrom the analytical trap is equivalent to 0.03& of exhaust/y£ of solution.

The TOE samples obtained from the preparative traps, which represent afairly large volume of exhaust, are suitable for odor studies as well asthe total analytical methodology consisting of gas chromatographic (GC)analysis, analytical liquid chromatography/ultra-violet analysis (ALC/UV),preparative column chromatography for the isolation of the aromatics (LCA)and oxygenates (LCO) fractions, and high-resolution mass spectrometry(HRMS) analysis. The small volume of TOE sample from the analytical orpreparative sampler is sufficient only for ALC/UV analysis and GC analysis,

* See Figure 1 for process flow.- 18 -


IEC

1X

tc.

- 19 -

Arthur D Little Inc

Analytical Methods

An approximation of the total hydrocarbon content of the TOE samples canbe made by low-resolution gas chromatography analysis, i.e., a ballisticprogram from ambient temperature to 205°C on a short 10% OV-1 column,46.0 cm x 0.35 cm stainless steel. (A Perkin-Elmer Model 900 GC unitwith a flame ionization detector was used for our work.) A 1% solutionof fuel oil has been used as the calibration standard. The same modecan be used to analyze the aromatic and oxygenate fractions to determinetheir relative mass content. This analysis gives the ug abundance ofeach fraction per £ of exhaust.

Analytical liquid chromatography - The ALC method is based on the useof newly available commercial components assembled as shown schematicallyin Figure 9. The following components are used:

(1) Two solvent reservoirs, hydrocarbon and alcohol

(2) Positive displacement pump

(3) A silicon septum sealed injector

(4) A 30-cm x 0.6-cm O.D. glass column packed with asilica gel-type support-Corasil II (Waters Associates,37-50 microns)

(5) Alcohol solvent injection valve (50 \ii)

(6) Ultraviolet absorbance detector and amplifier,detecting wavelength 254 nm

(7) Recorder

The unit operates essentially the same as a gas chromatograph, exceptin the liquid phase. A solvent flow is established through the column(1 m£/min), a sample is injected with a syringe, and the separatedcomponents which absorb ultraviolet light at 254 nm are recorded asthey elute from the column.

The aromatics (LCA) are eluted with the solvent front in the 35% CH CL /hexane and the oxygenates (LCD) are eluted after injection of 50 y£of methyl alcohol. The resulting chromatogram is shown in Figure 10.

The aromatics (LCA of Figure 10) represent the oily kerosene odor notes,while the oxygenates (LCD) represent the smoky-burnt character.

It was established that optical density varies linearly with sampleconcentration for both aromatics and oxygenate peaks. Thus, it becomesa simple matter to determine the aromatic and oxygenate content of everyexhaust sample by injecting 10 p£ aliquots (ca equivalent to 0.3 to 0.5of exhaust) of TOE samples.

- 20 -Arthur D Little, Inc

c

<3

o0

X

B0

\

| r

OlcC .0o o>5 5"i ? -< § » !I *(75

f

0)««5-,2

s.\

*— O

c

IV^H

s•M

7

ic

3

,-j" '"n i

"c5 "re

^s-g4) Q>£ S^O O

1 I< GO

^ SE w

S Q«N

E « fe

P'ct <» '

ocCD

I

iSBD-i o

6 <84) to

I"

CO1110.

OCCO.

Occ<111COUJcc

UJcc

8eg

I

- 21 -

Jet exhaust: 0*34

'LLMeOH

TIME Min.

FIGURE 10 ALC/UV CHROMATOGRAM OF JET EXHAUST TOE SAMPLE

- 22 -

Arthur D Little Inc

Preparative scale liquid column chromatography - Large volumes (500 £)of exhaust can be resolved into the three characteristic fractions —paraffins, aromatics, and oxygenates — by adsorption chromatography onsilica gel. These three separated fractions can be submitted for HEMSanalysis or odor studies.

A glass column, 30 cm x 1 cm I.D., is dry-packed with 6.5 gm of freshlyactivated silica gel (60-200 mesh, Grade 950). Pentane (distilled chroma-tographic grade) is passed through the packed column until the silica bedis homogeneous and free of trapped air bubbles. .Then the TOE sample(in 10 m£ of pentane) is added to the column and eluted as shown in Table 3to give the desired fractions.

TABLE 3

PREPARATIVE COLUMN CHROMATOGRAPHY OF EXHAUST SAMPLES

Components OdorFraction Solvent + Effluent Volume Eluted Characteristic

LCP Pentane, 25 m£ Paraffins Odorless

LCA Methylene chloride, 10 m£ Aromatics Oily kerosene

LCO Methanol, 25% > in o Oxygenates Smoky burnt

Methylene chloride, 75%

Mass spectrometry - To better define the chemistry of the LCA and LCOfractions derived from preparative-scale separation, it was possibleto make use of the analytical, capabilities of mass spectrometry in eitherthe low - (for LCA) or high - (for LCO) resolution modes. During thestudies of diesel exhaust,3we were able to show that matrix analysisof the low-resolution mass spectrum of the.LCA fraction containing theoily-kerosene odor would provide a measure of the amount of alkyl benzenesand indans/tetralins in that sample, as well as the more abundant naphtha-lenes. Further, an indication was obtained that the amount of indans/tetralins injected into the odor test room, as computed from the analyticaldata, did relate to the kerosene odor intensity. To a lesser extent, thealkyl-benzene concentration also correlated with the oily odor intensityof those fractions.

We also suggested that, although the oxygenates' smoky-burnt odor fractionwas much more complicated, a comparable type of analysis might be possibleutilizing data obtained from the complete high-resolution mass spectrumof the LCO liquid chromatography fraction. It is important to rememberthat we do not know with the certainty that was established for theoily-kerosene odor fraction just which groups of oxygenated species weprefer to measure to reflect the odor of that sample. Therefore, an

- 23 -


analytical method for the oxygenate fraction should summarize all of thechemical species concentration data where possible, but leave a maximum,possibility for examining the data in several combinations.

The oxygenates1 analysis then is a means of representing, on an internallyconsistent basis, the different types of chemical groups present in theLCO fraction which were identified as odorous species. The matrix(Table 4) does not represent the actual percent concentration of thesegroups in the sample, but rather a relative abundance influenced bydifferent instrument response to different species. The system doesprovide an efficient means of comparing samples and searching for odor-significant differences. The R+DB value is a definition of chemicalclass based on hydrogen unsaturation, while columns Oj and 02 indicatethe relative amounts of each class found with that number of oxygens.Thus, RfDB 1/0i could be aldehydes and ketones, R+DB 2/Qi unsaturatedaldehydes, R+DB' 4/Oj phenols, R+DB 6/02 hydroxy indanones, etc.

Combustor Test Program

Preliminary combustor experiments were carried out with inlet airparameters approximating an idle condition and varying the fuel/airratio. In order to determine the effects and possible interactionsamong these parameters on odor intensity and character, all combinationsof the four operating parameters were scheduled as test points. Theprescribed conditions were as follows:

Inlet air temperature °C 40 240

Inlet air pressure atm. 1 4

Air reference velocity m/sec. 15 30

Fuel/air ratio .008 .016

Of the 16 possible test points, we were unable to operate the combustorat six of the conditions,'primarily because of low fuel flows and poorfuel atomization.

The teat points selected for studying the five fuel variables and thethree combuator design variables could be limited to two (or three)cases. These were as follows:

Inlet air temperature °C 150 150 240

Inlet a i r pressure atm. 2 2 4

Air reference velocity m/sec. 15 15 15

Fuel/air ratio .008 .016 .014

- 24 -


• TABLE 4

HIGH-RESOLUTION MASS SPECTROMETRIC

ANALYSIS OF

OXYGENATE FRACTIONS (LCO)

(where ASTM-A1 fuel was burned)

R&DB

Numberof

Oxygen Atoms

1

2

3

4

5

6

7

8

1

16

12

11

9

9

5

4

1

6

5

3

4

5

3

2

Numberof

Oxygen Atoms

1

18

19

9

15

11

3

2

0

5

3

2

3

2

1

1

Mass

Odor

CEFF

8

1.3 TIA

99.0%

33

1.95 TIA

90.0%

- 25 -

Arthur D Little, Inc'

Because of the unexpectedly high efficiencies, additional tests werecarried out with the inlet air temperatures reduced to 95 C and10-40 C (ambient air temperature).

Each test point was maintained for 60-90 minutes to allow for samplecollection and odor analyses. The observed operating conditions andexhaust analyses were summarized as the average of 2-6 test pointsrecorded in the data acquisition system. With few exceptions, itwas possible to complete three test conditions on each test day andgenerally each condition was replicated on the second test day. Theorder of presentation was not randomized, but was dictated by thefacility scheduling and operational convenience.


IV. RESULTS

As indicated in the Introduction, the experimental studies are dividedinto three groups of experiments. The first group of studies was todemonstrate the applicability of the techniques developed in the dieselexhaust program to the analysis of jet combustor exhaust. The secondgroup of experiments was an odor survey on the effect of combustor oper-ating parameters on the odor level and character of the exhaust. Thethird group of experiments was conducted primarily to determine theeffect of combustor modifications and fuel variables on the odor of theexhaust. The results from each group of experiments in the programlogically proceeded to the next set of experiments with assurance of ahigh probability of success.

Preliminary Testing

Early in the program, it was deemed essential to demonstrate the val-idity of using the methods developed for collection and analysis ofdiesel exhaust for sampling and odor/chemical analysis of jet combustorexhaust. Therefore, two test points were sampled prior to the install-ation of the on-line monitoring equipment. These two conditions approx-imated the idle condition for a jet engine and differed only in thefuel/air ratio. Duplicate preparative-scale collections were made ateach test point and taken to Cambridge for extraction and analysis bythe ALC. Preparative-scale separation was carried out by liquid chroma-tography to provide the LCO fraction for high-resolution mass spectro-metric (HRMS) analysis. Odor studies were carried out with the TOE.These analytical results are found in Table 5.

TABLE 5

OPERATING CONDITIONS AND EXHAUST ANALYSES

Exhaust ComponentsTOE LCA LCO

(yg/A) (yig/&)

Operating Conditions

Sample Code

NA-1

NA-2

NA-3

NA-4

Tl(°0

150

150

Vr

(m/sec)

15

15

Pi(atm. abs.)

2

2

f/a

.008

.013

180 22 25

196 29 24

81 10 14

. 64 13 15

- 27 -


The TOE abundance was higher with the lower fuel/air ratio averaging 190pg/& at 0.008, and 72 pg/£ at 0.013. Similarly, differences were foundin the LCA and the LCO fractions with particularly good agreement betweenthe two duplicate samples.

A larger aliquot of the TOE was separated by preparative-scale liquidchromatography to provide an LCO fraction for HRMS analysis. From thisanalysis, it appeared that the relative distribution of the species

containing 1, 2 or 3 oxygen atoms per molecule, and 1 to 8 degrees ofhydrogen unsaturation was approximately the same as that found in similaranalyses of diesel exhaust extracts. No unique species appeared fromthe jet combustor samples and, indeed, there were fewer 03 speciesthroughout the set. This finding gave us confidence that the chemistryof jet combustor exhaust was not unique and, therefore, the detailedanalytical information obtained during -three years of work with dieselexhaust was directly applicable.

Preliminary odor studies were carried out in the static odor test roomin Cambridge at exhaust aliquot concentrations routinely used for thediesel studies, i.e., 20 liters of exhaust in a test room volume of12,000 liters, a dilution of 600/1, or a concentration of 1.6 £/m3 ofdilution air. The results of these odor studies indicate a highertotal intensity of aroma (TIA) for the fraction obtained from the ex-haust at 0.008 fuel/air ratio than for the 0.013 fuel/air ratio. Bothexhaust samples were noticeably higher in TIA than the reported valuesfor the diesel exhaust at a comparable dilution. Also it was notedthat the panel did not generally describe the presence of kerosene (fuel-related) aromatics usually associated with the diesel exhaust. This isconsistent with the finding that the LCA fraction was indeed lower thannormally found with diesel exhaust.

Odor Survey of the Effects ofCombustor Operating Parameters

This group of experiments had the dual function of training the newNASA odor panel and determining those operating parameters which pro-duced recognizable changes in the odor intensity, in order to simplifythe test program for the final group of experiments. The test schedulewas set up to examine the two extremes for each operating parameter(T., .V , P., f/a) in all combinations. This made a total of 16 testpoints which were scheduled for four test days, i.e., four test pointsper day. This proved to be an ambitious objective and actually onlythree test points were attained in each of the test days. However,several of the proposed test points were not feasible to operate and,as a result, 12 tests were carried out which represented 11 differenttest conditions. These results appear in Table 6, with the operatingconditions, odor intensity for the ordered presentation, efficiency,

- 28 -


w,-1

COwto

gHW

U

COzo

QaoCJ

oaM

3

w

N]

i

X

PM

Q

,-JW

CO4-)ccuCO&e0

CJ

4JCO3cfl,G

£

>•oG

•HCJ

II iMH

MHW

4-1•HCOc4-1aj_)

M0

T3O

COGO•H4J•HT3

CJ

M

5

COMQJP.O

°*lO "--JCJ onj_4 _^|

°^ 1^3 ^-JC-? ODh-j 3

c^C^ *x«^j &0nq

FT.rH

w

CO^B

|_J ,-..,>

H -H

<SJ

OrH 0

COI— I ^s^QJ bC3

CO^-^4H

B•H 4J

PH CO

CJrH 0)

> CO^s^

E

•H CJH 0

4-1 atCO T30) OH U

~* vO -3-vo -vt m

-* CM t-«co in coi-H

0 0 OO O l-~rH rH COCO rH

rH 00 O

•H r>. r»r oo <TI

mm o r»rH CM iH

rH rH CM

^ct- <t oo

ao oo vo0 0 rH

o o o

.-3- -* -si-

m vo voI— 1 I— 1 rH

o in mcO **^ *s^1

•H ' CM COO O OrH i-H rH

*

* -r- ,

00corH

OoCOCM

CM

COr--.

oCM

oorH

COI— 1

O

rH

^3-

rH

mCO

s^1

0i-H

* -OI— I

vO

rH

O

O

VO

CM

r~

o\

moCM

COCM

vOrH

O

rH

-strH

inCO

m0r-H

r * cj> oo CM in oo ••st*in CO VO CM I — 1 rHI-H rH

vO CT* VO rH VO rH VO

VO CTl rH VO CM rH •— 1rH CM rH

o o in m m o CMO O I

1*. 00 CO rH CM

co co -st* cy>CM CO

*xf LO oo r*^ ^3 T*** r*^

CN r * vo ^" o^ o^ oo00 \O Q\ 00 O^ O^ O^

m in inrH C^ C^ CT\ CO -s^ CM

CM rH rH rH •— 1 i-H rH

-sT CT> ^s, sj O O r-s•-H in CM •— i O m CM

rH

in CM rs. CTI vo vo cy>rH rH rH O i— 1 •— ' O

o o o o o o o

inrH rH rH rH CO «sf -it

r * r oo ^* r tn inrH CO CM CM CO rH •— 1

o in m o o m mvO CM O O CO CM CMrH rH CM CM CM CM

vO t~s 00 CT> O rH CMO O O O rH rH rH

rH rH rH I — 1 rH rH rH

CO

•H4-1cflH01

OO

0)CO3

CO0)H

eo

COQ

- 29 -


and exhaust analyses listed. Odor analyses were carried out at fiveconcentrations in duplicate presentations, first ordered and then random.Analytical samples were collected at all test points in duplicate withthe exception of those test points in which a preparative sample wascollected. The results from Test 101, the first experiment, were excludedfrom the analysis of the data since the panelists' responses were incon-sistent.

The exhaust hydrocarbon levels varied from over 3000 yg/£ for the lowefficiency condition to 10 pg/& for the highest efficiency operatingcondition. The LCA ranged from a high 300 pg/fc to a low of 11 yg/£,generally in a similar relationship with the total hydrocarbons. TheLCO varied from a high of 160 yg/& to a low of 8 yg/£. The three setsof data were generally parallel in these experiments. The data can beseparated into two levels of intensity and two levels of LCO. Those inthe range of 2 TIA units are associated with values 40 yg/£ of LCO orabove and those values below 1.5 TIA units are associated with levelsof LCO of 10 yg/£ or less. The low odor intensity values are associatedwith the efficiencies of about 99%.

The set of data can be summarized by plotting the total intensity ofaroma (TIA) vs the log of the LCO concentration in yg/£ for each testpoint. These data appear in Figure 11. The least-squares line forthese data is computed as TIA = 0.73 + 0.64 logCLCQ. The analysis ofthe data indicates that a straight line is indeed the best fit, withno quadratic, cubic, or quartic components of significance. Surprisingly,the variability (as indicated by 2a) is + 0.25 TIA units or significantlyless than the scale measure used by the panel to record intensity units.

Semi-log correlations are also observed between TIA and concentrations ofLCA and THC. The respective curves are TIA = 0.77 + 0.57 logCLCA andTIA = 0.4.8 + 0.47 logC-jHC' In b°tn instances, the slopes are lower thanthe relationship with LCO. The importance of this factor is the observa-tion that, in the dose/response relationship between TIA and exhaust,the values tend to fall between 0.8 and 1.5 with the majority of theslopes above 1.0. If the correlation is to hold that LCO is a measureof TIA, then the slope of this analytical measure should approximatethe slope of the exhaust curve.

Quite independently, but also of importance, is the fact that when LCAand LCO are separated by liquid chromatography and presented to odorpanelists in Cambridge, the TIA and the character of the LCO fractionmost closely approximates the values obtained for the exhaust (or TOEwhich represents the exhaust). Although LCA has odor at the concentra-tions examined, its intensity does not apparently influence, the TIAunless the concentration is as much as 10 times the value of LCO.

Of most importance from this group of experiments is the observationthat the panel can only differentiate two sets of inlet-operating con-ditions; those represented by the high-efficiency operation, i.e.,T of 240°C and P. of 4 atm., as compared with the inlet air conditions

- 30 -


oo

NlMO•a

Q

w

W

c/58 Drfi-J O

Q

g

003.

CO

a

VII

- 31 -


representing low efficiency idle, i.e., T. of 150 C and P^ of 2 atm.Therefore, only these two test conditions had to be represented in com-parisons of nozzle modifications and fuel variations in further studies.

The second important finding was the fine performance of the NASA panelduring this period.

Effects of Combustor Modificationand Operating Variables

The third group of experiments was divided into two sets: .thoserelating to combustor modifications and those relating to fuel variables.The analyses of these sets were treated separately. Two nozzle vari-ations were compared with the performance of the standard Duplex nozzleat three operating conditions.

These conditions were idle and high-efficiency test points, as wellas an intermediary condition, which from the original experiment waspossibly different in odor intensity from idle. The intermediary con-dition was constituted by an increase in the fuel/air ratio from 0.008to 0.016. This proved to be fortuitous since it provided some interestingchanges in the odor performance. In addition, with the Duplex nozzlethe inlet temperature was studied at two lower temperatures (approximately30° and 100°C).

Duplex nozzle

The operating conditions, odor and analytical data, for the Duplexnozzle appear in Table 7. As expected, the values for the high effic-iency condition produced TIA's below 1.5 and LCO levels below 10 yg/£,while all other conditions produced TLA values in the range of 1.6 to2.0 and LCO levels from 16 yg/il to a high of 80 yg/£. These values wereconsistent with July experiments, though there is generally a lower levelof LCO produced, which is also reflected in the slightly higher efficiencyof operation and the lower TIA values.

The surprising result was observed in Test 208 in which the odor levelwas essentially as high for the 0.016 fuel/air ratio as the valuesrecorded for the 0.008 condition and the emission level was significantlyhigher at 80 yg/£ of LCO. Operating difficulties encountered duringone test day did not permit this condition to be replicated. At thisintermediate condition, it was observed that the fuel flow had beensplit between the primary and secondary nozzle orifices. When the exper-iment was replicated with all of the fuel flowing through the primarynozzle at the same inlet air condition, the effect was to substantially

- 32 -


r»»

M_1

tq<3H

COwto[H

Ht/3

^. pqsPu

Q<tt>B0nHMQ&0O

O!2HIH

M

PH0

WidNCM055XWHHPH

- s""hJH|— ife

<|

^JEHw<U

1

co4-1c0)f>

o1o4-1CO

n)43XM

>!o<u-H

•rlmM-lf^l

>>4-1•rl

CO(30)4-1CH

MO

•O0

CO13O

••J•rt4J•HT3C

O

M13•H4-1rrtcuM0)AO

°<loSu onj 3

oil

d^d•j 3

«^1M MH 3i-tl

P4r_H4

HO

COa<u~S.M o?H rH

<S>

S0rH 0CX4 0)

COrH ~^0) M3 • •

EK

^14-1

B•H 4-«

PH cd

O1-4 CO

> co

~e

•H UH 0

v o o o i n o o - ^ r ^ . o r - ' v o•— 1 i— 1 CO -* CM CM i— i 00

i n c o r - » o o c M o o r ^ m c oc o i n < - i n c M ~ * c M c M /i-H T-H i-H

o o o o o o o o m ov o o > — i r ^ i n v o i — i o m < rr - » c o - d - < - i - H c o - * c o

t-H ^H

r ~ < i - c o c o o - * i n o c M i n• • • • • • • • • •

i n v O C O C M O O C O C O c N C T \ C T iC O O O C T l C r i C T < C ^ C y i C T v C T i C J \

in in in m "-1 m inv O O O r ^ O O vO O 00 00 '«* CO

• • • • • • • • • •r H r H i - H i - H i - H C M r H i - H i - H i - H

o o o o o o - ^ - i n v o c M o oc M c M - ^ r H c o i — i i - H c o i n m

C X 3 0 0 0 0 0 0 v O O O O O v O v O \ OO O O O r H O O r H i - H i - Ho o o o o o o o o o

,C N C S I C M C M C M C M C M C M - 3 - - 3 -

.

^ • ^ r v o v o v o v o r ^ p ^ i n mi - < i — l i — I i - H i — l i — i r H I — ( r H i - H

I O'n o o in m o mo o- ^ H < N o o > C T \ i n i n - j - 3 - ~ 3 -

r-l i-H CSI CN

4J CDCO T30) OH O

o o o o o o o o o ^C N 1 C S I C N C M C M C N C N C N C N C S I

- 33 -


reduce the odor level to 1.2 TIA units and to reduce the LCD levels toan average of 16 yg/£. These data are tabulated in Table 8. The odorintensities for the idle condition are somewhat lower (1.5) than thosepreviously observed (Tests 206 and 207), although the LCO levels areapproximately the same. The value for the high-efficiency level issomewhat higher (1.4) than might be expected for the value of LCOobserved, (8 yg/Jl). These data are apparently inconsistent in as muchas lower intensity of aroma was observed for the intermediate test con-dition with higher LCO values as compared with the high-efficiency testcondition.

Air assist nozzle

The first nozzle modification examined was produced by operating theDuplex nozzle with air assist on the alternate orifice. The data forthese tests appear in Table 9. The fuel entered through the primaryorifice at the two low-temperature conditions, and through the secondaryorifice at the high inlet conditions. In all cases, the odor levelwas below 1.5 and the values for LCO were in the range of 3-6 pg/£ ofexhaust. The levels for LCA were also in that range and generally lowerthan in previous data sets. Total hydrocarbon level was also reducedto 50 yg/&, or .below, and all of the efficiencies were measured asbeing 99% or above.

This appears to be a significant improvement in the odorous levelsfrom the jet comhustor.

Simplex nozzle

Based on the assumption that spray pattern, or droplet size, couldhave a significant influence on the odor intensity of the exhausta Simplex fuel nozzle was installed in place of the duplex nozzle.The data for these tests appear in Table 10. The orifice size wasselected to accomodate the highest fuel flow with the available pumppressure, and, therefore, was oversized for the low fuel flow, withthe result that high TIA's averaging 2.1, were observed for the idleconditions. These values are at least as high, if not higher, thanfor any Duplex nozzle operation, however, the LCO values were 30-40ug/£. The levels at the high efficiency were in the range of 2 to6 yg/&, and the intensities of aroma below 1.0. Interestingly theintermediate condition was also found to be relatively low in TIAalthough the levels of 14-16 yg/£ of LCO were consistent with thoseobserved for the Duplex nozzle operation with the full flow throughthe primary orifice.

- 34 -


co4Jacuo§<oo

w

0o

3I00CM

00

o00vO

CNCO

oo oo

O00

oOi

oo

COcd

COwC/3

oaCD•HO

w

FnsO oo

CO•o oo 00

CTi

60

O

CO•rl

00

w

OaX!W

,-J63g

oCJ

O

O13O

ro

<!HH

in•inCM

CO(5O

•Hj l+J

•HXICJO

CJ

00(3•rl4JcOt iM

a)ft0

grH Ofe CU

CO

cu oc3fn .

vO1-4

cfl-•^M-l

oo0o•

e•rl 4J

PM toCM

orl CU

> CO

~B

VO1 — 1

m

oooo

CSJ

CO-a-

inf—io

CM

. 'H uH O

4J CUCO TJCU OH CJ

m

om

Om

CMOm

m-a-

COom

Om

om

vO

o<rCM

mom

cu3

M-l

-cfOm

CO

COom

ft01oKO)

COco

•H13COU

COcu•H •t-l 00cu oCO CM

o jio *JCM -H

O T3CU

•U JJcO cdcu &ft 0CU O

P4 O

- 35 -


W

W

rfNJO

HCOHco

W

COO

onents

Q.

§

O

4JCO3CO

•aW

>iCJG0)•H0

•H14-1It i*M

W

>i4-1,j•n

COC0)4JC

t 1Mo

T30

COCo

>r4-j•HT3COoMC

•H4-1fltwH0)cxo

°MO -~~J co^ d

d^ N•J 3

0?|

0 -J 0a o3 coH 3.1

CufnWO

COe

^^r ^4,H H

CHJ

>o•-I 0IV. fl\

HH UJCO

^J "*•*!..

CU 003

fe .

O•en

CTv

mCN

i— i

i-»r

cO

MH"

oooo

•

&•rl 4J

PL, «J

oH 0)

> co"0"

•CN

r^

•H Ol 0H 0-1 «n

-1

4-1 0)CO t3Q) OH 0

.-iO-fr

CN vO vO vO

Oro in

OCO

OCO

CO•

o\Ox

CO•

CTiO\

CN

CTi

COmCN

inCO

mo

COCO

coco

CTioo

CN

vO

OvO

CNO

CN

COi-Ho

CN

CO1—Io

CN CN

VO

ovO

COO

vO

o-*CN

ino

vO

in-3-CN

vOo

- 36 -

Arthur DLitttelnc

CcuCO

£oa4-1CO

c«J3X

OU

0

co

mr--

OvO

ooro

Ovoco

CMCM CM

mm

CN

C/3

W

oCOOO

VOoo

CTi•

oo

oo

01— 1

fl]l_3

S3H

HWP

3XwQ£23

C/3zoMH

QaoU

CJsH

^WP-iO

wI_JNlNo

Xui-4PM

1— |

^,_^

W-^D

-i— i^3I5q§-

^

i4J

tnCcu4->CH

M0

0

COB

M o*H -H

<H>

m mCM CT.• •

CM ^H

COCO

•H4J•H""UCO

£0

rH CJfe (1)

CO.H ~-0) bC3

vO vO*-t i-l

cfl* s.

m

oo ooo oo o• •

QJ&O

FM

H O

m

CMCO

CM

vO

m

CMCO

vD•—Io

CM

vO

m mCOCM

incoCN

4J CUco -oco oH O

CMo

COo

mo

vOo

- 37 -


Fuel Variables

Four fuels, in addition to ASTM-Al fuel, were examined in the course ofthe program: natural gas with no aromatics and a gas at ambient tem-peratures, isooctane, a hydrocarbon, with essentially no aromatics,JP-4 with slightly higher aromatics, and JP-5 type, a specially blendedfuel, an aromatics concentration of 30%. The 50% boiling points (100%for isooctane) are listed below:

ASTM-Al - 205°CJP-5 type - 230°CJP-4 - 150°CIso-octane - 116°C (B.P.)

Each of the four fuels was examined at the three standard test points,and each test replicated on the second day. With the exception of thenatural gas, an additional test point was examined in which the inletair temperature was reduced to between 10 and 50°C.

Natural gas

As might be expected, the natural gas gave the lowest total intensity ofaroma. As shown in Table 11, all TIA values were below 0.7 and in thehighest efficiency condition in one of the tests, the value was calcu-lated to be 0.1, which is slightly below the value found for backgroundodor level of the test room, although the character was recognizable asburnt. Indeed, in order to measure the intensity of aroma from thisseries of samples, the concentration range was increased to a maximum,i.e., a dilution of only 35/1 at the highest level. This would suggestrecording TIA for 10£ of exhaust per m3 which indeed eliminates anybackground. Values for LCO were measured in all samples analyzed. Theinitial test results were apparently inflated by contamination fromprevious test runs. In other cases, values from 1.3 to 0.5 yg/Jl of LCOwere measured. These were generally commensurate with the intensitiesof aroma observed in this series of samples, although they are near thethreshold for measurement with the analytical system and standard samplesize used in these experiments.

Isooctane

Isooctane also produced uniformly low odors at all of the three standardtest conditions with the values for TIA near 1.0 and LCO values of 3-4yg/&. The exception was the lowest temperature condition, air inlettemperature of 10°C, where the TIA observed as 1.6, and the LCO valuewas increased to 10 yg/£. These data appear in Table 12.

- 38 -


CO4JGcuGO

I4Oo

*0o

033H

m•

oOO •K

•K

4C

*

oCM

O\OCM.

towto

3<5^J

Htop

wp<3to

oI—IHHP

0U

Cj)£5HH

pqwo

to

P^Hq

M

§

3H

MP-itoj

1

5o -dW£3

53

OfiCU•HO•rl14-1

S

4J•HWGCU4-1

cH

O•ao

fefeWu

CM OO CM O 00 00• • • • • •

LO ^o r** ' ' oo • cy\ c^cr> ON (Ti cr\ <y\ &

CO_S_

MH

COco•rl

•rlT3GO

oflG

•H4JcdCUftO

u

o

rHcu

Pn .

cdm

•rlPM

Mt>

•rlH

4JcncuH

o$

•-1

<2J

•TlCU1 i

m in m ^• • • • • • t.

4J•

CO

rH*

^

OcuCO

^s^

60

r•H4J

Oflj

vO tO r— 1 f-H Q O\ Xr— 1 i-H f^ O^ -^ CO gj

Q)^

••x^

oo oo m »o co CM °°O O r — 1 r H r - H r H ^ 1

0 0 0 0 0 0• • • • • » '"*

•

oB4Jcd

CM CM CM CM <t *& *%M^u

OcuCO

~>^

oo•

or*» - LO r^ \o u"i to ..i-H " «-H f-H t-H i—H ""* ri

CO1— 1

u0 iO tO iO <O ^^ LA *°-^" *<f -d* -^ ^ en^H r- 4 ^-4 _J CM CM n

Q)1 i

CO>>

CO

cu

1 p— 4 CM CO •i " t-O \O0 ' O O O O O .,<?> CTS 01 cy> os cy> *

•Go•rlj l+J

rrttyriM

cd4JGOo

!CO

CO

,_!

cdoMft

CO4-10ODCO(1)

w

S4w

,__j

(U

^*ri

aa)P<end)H

o?

00p-

o•

cr^T)

cdo

•m

•K

*

•

CU4-1

Ocu4-1cu

Qi

4-1O

p

•

CUNi

Hcd

«Si

4-1oa

<a

- 39 -


CO <• <r co-* -*

CM <-< CM CM CM CM

O O O O CO O

CMiH

W,-J

EH

COW'cot«3

2HcogM£sW

Q

^CO&sHM

OCJ

C5

5HWwPHO

W_3

NJtSJOa

WJ_J

BpTJwgw .

HOooCOM

0a01•HCJ•H*4-414-1W

fVj

pnwCJ

CM CM ^H CO ^H CM• • • • • •

CT\ r^ r^ oo oo c^GO CT* . - CT\ C^ ON ON

^14J•H

COC0)4-1aM

M0

O

COB

H ^H

C2)

vO CT> O i-l 00 O• • • • • •I— 1 I— 1 •— 1 I— 1

COCO•rl4-1•rl*"OcooM-.5H)M0)P-0

OrH CJfn cy

CO

01 003

pL^

o\ -3" in vo r~ vo•-I •-! ^H CM CM CO

n)* *«.

4-1

oo co co m m CMo o o >-i •— ' •-<o o o o o o

• •. • • • * •

B•H 4J

P-i COCM CM CM CM ' CM -*

CJH OJ

•H CJH O

•^ \o vO- vO vO ' • vO

o o m m m "^

*—t i— 1 .— 1 i— I CM

<••

O"*o>

o•

I— 1

vOCO

COi— iO

•

"*

m

mcoCS|

CMOVO

COOvO

OvO

inovO

vOovO

ovO

- 40 -

Arthur D Little Inc

JP-4

With JP-4, TIA values for all of the standard conditions fell in therange of 1.4-1.6 TIA units. A possible anomaly in the data appeared atthe intermediate condition where the concentration of LCO increased fourfold to 35-38 yg/X,, with no measured change in the TIA. Curiously, atthe low inlet air condition, 40°C, the TIA in both instances was 1.8,with an LCO level of approximately 13 pg/&. These data appear in Table13.

JP-5 type

The data for JP-5 type fuel appears in Table 14. The odor levelsfrom JP-5 were somewhat higher in all instances than the values ob-served with other fuels, including ASTM-A1 at the standard operatingcondition, while the LCO measurements were in the range of 5 to 10Vig/fc. The one exception was the high efficiency test condition wherethe run was aborted midway through. This may have influenced the hydro-carbon level during part of the collection. TIA values for the ineffic-ient condition were 2.0, but were associated with only 10 yg/& of LCO.

An interesting observation here is that the least-squares line for thesevalues approximates quite well the line determined for diesel exhaust.(TIA - 1.1 + 0.9 logCLCO vs TIA = 1.0 + 1.0 logCLCO for diesel).Indeed this fuel was designed to represent a commercial-type diesel fuel.With the low inlet air temperature, the LCO value increased to 24 pg/£,while the TIA measure was 1.8 which is consistent with the values mea-sured for the other fuels at this reduced efficiency condition.

Odor Character and Composition Variation

The primary odor characteristic of all of the exhaust samples is asmoky or burnt-smoky odor. With the low aromatic fuels the odor typeis more sweet and burnt rather than smoky and tarry that is perceivedwith the higher aromatic fuels. In general, there is little fuel-related character in combustor exhaust. When it appeared it wasusually associated with low efficiency conditions and most frequentlywith JP-5 and JP-4. A sdlventy odor, characteristic of the isooctaneitself, was detected in the exhaust produced with that fuel. Odorprofile composite descriptions of the exhaust from the five fuels appearin Table 15. Pungency or with JP-4, nose irritation was apparent in allsamples. An oily note was perceived in all samples but was describedas oxidized oil only in those fuels that contained significant levelsof aromatics.

- 41 -


cowCO

1Hen£3jpH

W

Q

Ê

^Q53O

HM

O

O£5HH

$WPHO

CO4-1ccu

. ao

!u4JCO9n)

rCXw

od0)•Ho•H»4HII i*W

w

°^ 1O ~--.l - 3 - c o c s i o o m o o o o o oO Oq rH rH rH CO CO

nJ nl

< J ^-^1 m c o v o < f i n c s i r » » v oc j t j q c T i oo I-H —i vfj rîJ J

o < | O O O O O O O OO ^^-J vO O r^ vO CT\ OO O> C^p S o d r H oo CN csi vo r^8^ — 1 1 rH

fnpT|

Wu

O O\ rH C O C S ^- CTi VO• • • • • • • •

c T t O v o v o i n * 3 " o o o o00 CT\ CTN tT* CT\ C7N OS CJS

W £*,I-H 4JN -HN] CO0 C23 CD

4-Jt^t f*t^* W

W M

5 on -a-- o

COg

M ^~H 2

<3>

in oo o o o c o i o v o - * i n s r1 — I r H r H i — l l — I r H r H r H

w

1H^

COCO•rl4-1•HT3ao

M

•rl4JtflM0)

°

Or-l CJPC* Q)

COt— I *««0) M3

c y i C T i v O v o c s i c o c s i c s ir H r H r H r H C O C O - ^ - J -

cfl

U-l

o o o o o o o o v o m c o c oO O O O ^ H r H r H r H

o o o o - o o o o• • • • • • • . •

B•rl 4J

CM n)

'c s i c s i .csi c s i c s i c s i - a - - j

0M CU

> CO

•H aH O

4-1 CDCO T30) 0H O

•sf â* r * r r r * in inr H r H i — I t — l i — I r - I r H r H

o o m o o i n m orH < t - j - mm - a * co - a *

rH rH rH rH CM CSI

I-H <N co -a~ m vo r ^ o oo o o o o o o oc o c o c o c o c o c o c o c o

- 42 -


COvO

m CNIin

oom

oom

mo

moo

COw.CO

oo inCO ro

CTioo

w

HWS3

p-l

FsW

Q

<!

C/3552p^HHQa0u

81MP-4O

Cs]tNO55

XW,_jP-i

Q--i

H-f&3

£

*wPLI

£

inI

P-I1-3

H4-1•rt

CO(3014JaM

» ir*O

TJO

CO-s^^rM °<H ^

CH;

moo o cri r-» vo vo

• • • • • • .r-t CN *-1 I— 1 1— 1 t— 1

COCo

•H4-1•HT3GO

C_J

M

•SJ

COMa)PL&

i 0i-l OPLI <U

COrH0) M

f^ .

•

r-4 in m o cn CN coCN -H •-< m si- <r tJ

4-1n)eoM

CO

14-1

E•H J-i

P-I CO

oo P^ oo m -* <f ™O O O •'-> <-! '-Io o o o o O s-s. . . . • .0

m

iCN CN CN CM ~* <T

• . " r— 1w3

<4-l

OM 0)

> CO

B

T3^t- 1^3 \o vo m m cu1— 1 r— 1 i— 1 i— 1 I— 1 r-4 tJ

G• CU

•H UH 0

i — io m in o ""> o ,0en -* <r m ^ •*

r-l r-1 r-H CN CN >,1-1r-lca

* -H

4J 0)CO 130) OH <->

* 01-1 CN oo -j- m vo o)o o o o o o P-o o o o o o wT— * T— 1 ^H 1-H «— * t— < HC

•n— tTJcu4-1M0

43cO

4JCO0)H

.If

- 43 -


PM>-3

MHCOH

CM3

CJcIU

C0*

<4H C<4-l E»> 3

PQ <U <U rH PM^ fl -rl

CO !5 O 4->fs CO O J3

X O rH M 00O )-i N-' 3 -H8 CO O rH

co & to co

800

covO

OO

603.

oooo

w

PL,

g13 ^ iH0) M 4JN 0) CO

3 COco C

•H CO rl.. X CO HCO O rH M

•M CO v~'c o ^ corl rl rH CO' 0) -rl O

:*$ O 2"

8rJ

M.3.

vO

M

oCJ

003.

COVO

co Qo w§ gO M

W

^ £ 13* §§ 1O SB

1— 1<t!1aHCO<;

4-1cn3

pq

>,J4o0

CO

coCCO4-1OOoCO

I-H

>,

oaCO

4-1

g

3

CON•rlT3•rlM0

s*.rH•H0

>»rH•rl

aco>

rHO

CO

4JCOCOS

CO

4-1cco>rHO

4->CJCO00c3

PU

4JCCO0003

X"N

>*4JCOCflCU>>

V— X

M30

8003.

co

003-

OO

00

OCJ

003-

PQ O co PL, co

34-1CO3

4JCO(Us

CO

4-1ai-t

p<COoCO

PQ co O

g00§

PM

OCJ

003.

u-iCO

OcjnJ

003.

CO

•uao

orto

- 44 -


V. DISCUSSION OF RESULTS

As a result of the experimental studies, it is clear that inlet operatingparameters, fuel types and combustor modifications can all influence theintensity of exhaust odor calculated for a standard concentration ofexhaust in dilution air. Associated with the changes in odor intensityare changes in exhaust composition and combustor operating efficiency.The experimental results have been examined independently and in combina-tion to determine which of these measures, expressed logarithmetically,might be most useful for estimating the sensory response. An importantconsideration in selecting analytical measures for correlation with odorintensity is the relationship between the measured species and the sensoryresponse. For this reason, our work has emphasized the correlationbetween odor intensity and the concentration of the partial combustionproducts as measured by LCO.

Because of the practical value of a correlation between odor and operatingefficiency, this data for tests with ASTM Al fuel is included. Anexpression which represents inefficiency of combustor operation (100-CEFF)has been used to graphically present this relationship. This may also berepresented by the milligrams of LCO per gram of fuel burned.

Finally, in evaluating the environmental impact of combustor exhaust,the total quantity of odorous material emitted during a prescribed timedetermines the "odor load" on the diluting atmosphere. This is the.product of the odor level of the exhaust (LCO) and the volume of ex-haust emitted per unit time.

Odor Analytical Correlations

Including the data from all tests carried out under this program, theexpression for odor strength is found to be:

TIA = 0.68 + 0.72 logCLCQ

2o =0.46

r2 + 0.67

The correlation of TIA with LCA is at least as good for all experiments,except, of course, for those in which natural gas was the fuel since nomeasurable LCA was found.

- 45 -


The correlation can be improved by the inclusion of the on-line exhaustanalytical data (represented by CEFF & THC). Combustion inefficiency(100 -CEFF) gives a slightly more significant result than total hydro-carbon data. Including this measure in the regression analysis leadsto the expression;

TIA = 0.75 x 0.55 logCLCO + 0.23 log (100 -CEFF)

With this addition the correlation as measured by r2 is increased onlyto 0.72.

Comparing the correlations for the ordered presentations with the datafrom the random presentations, had the effect of reducing the TIA inter-cept by 0.1 TIA units. Since the inclusion of the random data did notaffect the conclusions and increased the variability, they have not beenused in the final correlations. Comparison of the analytical resultsfrom the prep scale samples and the analytical scale samples shows theprep scale to be about 15% higher on the average. However, the correla-tions of TIA with each set of samples independently are not significantlydifferent. Therefore, in the analyses all analytical values for each.test point have been averaged.

Combustor Variables

During the last phase of the program, odor data associated with exper-iments using ASTM-Al fuel and with variations in the nozzle operationswere compiled. They are summarized in Table 16. From Figure 12 it canbe seen that this spread in values extends the range of concentrationsstudied by one order of magnitude as a result of the exhaust from theair assist and Simplex nozzle modifications. The calculated line;

TIA = 0.79 + 0.60 logCLCO

2a = 0.46

r2 = 0.58

is almost exactly the same as that determined for the odor survey data.

It can be seen from the data that the air assist nozzle reduced boththe odor level and the concentration of oxygenates in all three testconditions, but most significantly in the low-efficiency conditionwhere there is an order of magnitude reduction which is associated witha 0.5 TIA unit decrease in the intensity of odor. The Simplex nozzlealso produced a significant reduction in the odor at the high fuel flowcondition, but was high in odor with low fuel flow. An interestingobservation which emerged from the first of experiments was the factthat in the intermediate test condition, 150°C - 0.016 fuel/air ratio,when the fuel flow was split between the primary and secondary nozzles,the efficiency was lower and the odor intensity as well as concentrationof oxygenates were substantially higher.

- 46 -


ou O

oo

§o<uen

oo

w

CO

H Q ^ P*

W <i § f*

WCO

HWi-J3

COpo

CO3OHH

CJ

HQ

i

X0)

r-(

4-1to

•Hto

• Jcj an-l \

CO

H c

f)

m o oi •<r -* m

cs

m m

m o o. • •

oi m • vo

CM m

m

X01

p.Q

O — • •o ad in ^o^ ^ ™

<"ai— i ~~- ^ ^• •

•oo

<r•

•vOr-H

CM•

•ps^

ro

00•

•0CM

0•

•p^i— 1

m•

CO

co•rl4-t

c0

CJ

4Jto0)H

rtm

00 vO

O O

•H CJH O

0 O

^H CM

<•^H

O

oCM

•— IO

0mt— i

oooo

mCT>

rH

o

mON

00oo

o

_ 47 _


frco cd

T3 -H -^ T3M CO X !>-, CCO CO Q) M O*& ^d *~i cct oC &, 0 cu(fl V< S -H c/>W -rl -H t-JW <J CO CM

CO§l-l

>

W

NO

3- O

o 8S

00

co

H

cs

s

VII

- 48 -


Fuel Variables

The second set of data from these experiments was the comparison of theodor levels from four different test conditions for five different fuelsreported in Table 17. It was apparent that as the volatility of thefuel increased from ASTM-A1 to natural gas, there was a concurrent re-duction in the intensity of odor at all of the exhaust conditions, withan associated reduction in the concentration of the oxygenate species.Natural gas, surprisingly, does produce odor but at very low intensitieswhen examined at 1000/1 dilutions. We were able to measure concentra-tions of the oxygenates in these samples. The reproducibility is notas good at these levels because of low LCO values and the interferenceof background. The correlation found between TLA and LCO when the datafor all five fuels at similar operating conditions were analyzed is:

TIA = 0.59 + 0.89 l°gCLCO

2a = 0.55

r2 = 0.74

This line has a higher slope and lower intercept than for all testswith ASTM-A1 fuel. The curve of TIA vs logCLco appears in Figure 13.When all of the data from all of the experiments is pooled, however,there does not appear to be a significant difference in the relation-ship between TIA and LCO for the several fuels.

If the TIA intercepts and the slopes for the individual fuels are com-^pared with the values for all fuels combined, one can use the "t" testto measure significance of the differences. No significant differencesare found between the various slopes. There is a 95% probability thatthe intercept for JP-5 (1.68), Isooctane (0.44), and natural gas (0.35)are significantly different from the ASTM-A1 value (0.73).

Efficiency and Emission Levels

In examining the other parameters which may relate to odor intensity,and which are normally measured in combustor tests, there appears to bea reasonable correlation between the TIA and the log of the inefficiencyof combustor operation, represented as 100 -CEFF, and plotted TIA vs itslog (Figure 14). Over all of the Task III ASTM-A1 fuel tests, it can beseen that the values fall in three separate groups with highest effic-iency observed with the air-assist nozzle operation, and the lowestefficiency observed with the lower inlet air temperature and low fuel/air ratio. The least-squares line for these values, the values rep-resenting all of the ASTM-Al experiments, is TIA = 1.3 + 0.77 log (100-CEFF). Thus at 99% efficiency, the intensity of odor is approximately1.3 TIA units at a 1000/1 dilution. When engine operating efficiencyis increased to 99.9% the expected intensity of aroma should be approx-imately 0.5. and indeed it was found to be 0.75 TIA units.

* Students "t" test for statistical significance.


rt oo--. om o

O

-H°0H si-

oo-I a

ro m

m•

co

00

o•

oCM

oo

w

HCMM

3wH

CO

gsoCJ

COMO

CO&sH1-4Q28

HWh-1&

CO

CO

o

•H°0

M-l O•

u-H%H sf

(0 00-- OIH O

O

•H°0

OO

CO

td

CO

0•

v£>i— I

CM•

r- 1

O

00

-*•t-H

<N

m

m o• •vO COCO

mm oi• •i— i

o o00 <!•

m<t 0• •

i-H •— 1

O 0

O <f

0 0• •

r-l VO

mm r^• •i— i

yCO O• •

rH Oi— 1

CM vO• •

r— 1

in oo

CJ\

co•H4J•HT3couwCO01H

r-l

<f<-4

rH COD <!3fe

cucto4-1CJ

<f O1 O

On CO•-3 M

CQcdO

r-Hnj^34Jctfz

mP.>.•U

mI

PMro


£

COCO

0) 00ao) iH4-J 0)

° ^O 3

H I O 4J ICM co ta p^>D M & *-i

x o D<

wgd<!

o=>? u60 ~--3 60

3.^^

COO >

oMfe

VII

- 51 -


Oo

W

CO

COHCOWH

O2W

W

H

st

O

VII

- 52 -


A similar correlation is found when the data for all fuels are combined.There is, however, considerably more scatter in the data. This wouldbe consistent with the observations that with different fuels differentlevels of LCO are produced and the hypothesis that the LCD's for differ-ent fuels have different values of TIA/yg.

It is apparent that there is also a relationship with efficiency and mgof LCO produced per gram of fuel (ASTM-A1) burned. At high efficiency(about 99% CEFF) the calculated values are 1.0 mg/LCO/g of fuel (0.1%) orless. The maximum value 10.5 mg/g fuel was found with a very low effic-iency of 67.5%, while the next highest value of 9.4 mg/g at 81.4% CEFF.The values calculated for a series of ASTM-A1 fuel tests are given inTable 18.

Although the milligram percent of odorant produced for each gram of fuelburned has an influence on the impact of jet combustor exhaust on theenvironment, the calculated total quantity of odorant emitted per unittime is probably more important. This is the product of the air flow(m3/sec) and the odorant load (mg LCO/m3).

The values cover two orders of magnitude from 600 mg/sec at low effic-iency and high fuel flow to 4 mg/sec at idle with the air-assist nozzle.

Even at the high efficiency where both air and fuel flows are high, theodorant load on the environment was found to be 87 mg/sec as comparedwith the average for several idle condition results of 50 mg/sec.

Odor Character and Composition Variation

Samples from natural gas and from iso-octane are recognizably differentin character from the other three fuels. It would also appear that theJP-5 exhaust could also be discriminated because of its sourness andburnt tarry character.

In correlating the odor intensities measured for different fuels withthe amount of LCO measured, there are significant (95% confidence) dif-ferences between the TIA per yg of LCO observed for ASTM-A1 and thoseobserved for natural gas or JP-5. In no cases are the slopes of thedose/response curves sufficiently different. The values are recordedin Table 15 as TIA found, that is the TIA per yg LCO from the observeddose/response curve and TIA calculated which assumes the same slope(0.89) for all fuels.

There do not appear to be significant differences in odor type betweenlow or high efficiency operating conditions for the same fuel. Severaltest points suggest that irritation factors are more apparent at the lowinlet air temperature (150 C and may be higher in strength at the higherfuel/air ratio.

- 53 -


VOO>

oo voco co m CM oo CM

CT\ vO(N ^H

m CMoo

fOCM

CM CNOO 00

CMOO

oopv.

oo in00 coCM CM

$ OT

oo

3ow wrJ PM ^

CO <JJ3 fe IO O H Senco

00<!

O i-(a sM-l

000 Ml

Oin oo

ooCO

COo CTi CM

00co a\

inoo

CO vOCM

CT>

wu

p

o

ooCMoo oo

vOOO

•*CO

v£) O CM

CO 00CTi O-i

ooen

oo

en0)H

CM CMo or~- CM

oCM

oin

O O

CM O vOO ^H O-d~ CM r*

O O O

- 54 -

Arthur DLjttklnc

The HRMS matrix analysis has been carried out for a total of 16 testpoints, the majority using ASTM-A1 fuel. Although differences appearbetween samples, they do not appear to occur consistently. Of moreimportance, however, is that data is not available for the two fuelswhich are most different in odor type and intensity from ASTM-A1 fuel.There was not sufficient material collected from the natural gas exper-iment for analysis and instrumental difficulties have delayed the re-sults for JP-5 fuel.

In addition, the complexity of the matrix, 16 independent variables,and the paucity of observations makes it impossible, to draw sound con-clusions. A principle components analysis8 was carried out for elevenmatrices, which do not indicate any significant relationships whichcould be used to simplify the data set. Multiple regression analyseswere carried out with TIA's as the dependent variable and although somecorrelations were observed, the coefficients were no better than withtotal LCO.

It is clear, however, from these data that the chemistry of combustorexhaust is similar in both types of species present and general chem-ical complexity to diesel exhaust. Detailed identification studieswould probably not produce more useful data than has been obtained.

- 55 -


VI. CONCLUSIONS

The experimental program showed that the odor intensity of jet combustorexhaust is apparently decreased by those inlet operating parameters andcombustor modifications which act to increase the burning efficiency ascomputed from the exhaust gas composition. The partially oxygenatedcomponents, as measured by an analytical liquid column chromatographicsystem, correlate well with the intensity of odor as determined by adefined method of the sensory response and potential environmental effects,regardless of fuel type or combustor operating conditions.

At comparable inlet operating conditions, the type of fuel burnt canhave a substantial effect on the perceived odor intensity. This effectis correlated with changes in the concentration of the oxygenates (LCO)determined in the exhaust. Similarly, those combustor modificationswhich significantly reduce the intensity of odor are those which reducedroplet size and increase the dispersion.

Changes in fuel type are associated primarily with differences 'in odorcharacter. The ALC system measures only the amount of LCA and LCO butgives no information on the composition of the LCA or LCO fraction.Other techniques such as high resolution mass spectrometry would berequired to furnish this information.

The sensory analyses by the newly-trained NASA panel have provided con-sistent measures of odor. In general, their response is somewhat lowerthan we have observed with the more experienced ADL panel, which mayreflect differences in sample presentation or less experience in intensitydiscrimination. It is possible, however, to obtain relevant odor infor-mation from sensory analyses of the total organic extract from an exper-ienced odor panel.

Recommendations

Since the present study has been limited to the exhaust of a singlecombustor at one sampling point, it is recommended that the experimentalstudy be extended to determine the effects of turbine operation, designvariations, and sampling points on the odor intensity and oxygenate*sconcentration. Sampling could be carried out at NASA-Lewis or atcommercial test stands and odor/chemical analyses carried out at ADL'sCambridge laboratories.

Because of the profound effect of factors which affect the rate offuel -vaporization on the odor intensity and concentration of oxygenates,we recommend an exploratory research investigation to examine in thelaboratory the effects of droplet size and particle distribution onodor intensity and LCO concentration in order to understand the poten-tial contribution of combustor design and fuel selection to the reductionof odorous air pollution.

- 56 -


The primary impact of combustor exhaust is in and around airports.We recommend that a study program be undertaken to attempt to measureodor intensities at various airport locations, and collect ambientsamples for laboratory odor and/or chemical analyses in order to esti-mate whether the predicted odor dilution values are realistic and todetermine the actual dilutions which can be expected under various atmos-pheric conditions.

- 57 -


REFERENCES

1. Anon.: ADL Report No. C-70131 and C-70132, Chemical Identifi-cation Of The Odor Components In Diesel Engine Exhaust, 1969.

2. Anon.: ADL Report No. C-71407-1 and C-71407-2, Chemical Identifi-cation Of The Odor Components In Diesel Engine Exhaust, 1970.

3. Anon.: ADL Report No. 62561-5, Chemical Identification Of The OdorComponents in Diesel Exhaust, 1971.

4. Anon.: ADL Report No. 73686-5, Analysis Of The Odorous CompoundsIn Diesel Exhaust, 1972.

5. Stevens, S. S.: To Honor Fechner And Repeal His Law: .Science,vol. 133, 1961, pp. 80.

6. Sjostrbm, L. B., Cairncross, S. E., and Caul, J. F., MethodologyOf The Flavor Profile, Food Technology, 1957, Vol. XI, No. 9,pp. 321-324.

7. Anon.: ADL Report No. 74744-3, Chemical Analysis Of Odor ComponentsIn Diesel Exhaust, 1973.

8. Miller, I., and Freund, J. E., Probability And Statistics ForEngineers, Prentice Hall, Inc., 1965.

- 58 -


DISTRIBUTION LIST FOR NASA-CR-121159 Arthur D. Little, Inc.

Mr. L. MagittTNaval Air Propulsion Test Center (AED)Philadelphia Naval BasePhiladelphia, PA 19112

Aerospace Research LaboratoryWright -Patter son ABB, OhioAttn: Dr. R. G. Dunn

NASA Scientific and Technical Information FacilityP.O.Box 33College Park, Maryland 207 0Attns Acquisitions Branch

FAA Headquarters800 Independence Avenue , S.W. . 'Washington, D.CJ'. 20533Attn: Library

William T. Westfield . .

NASA Headquarters600 Independence Avenue, S.W.Washington, D.Cv 205 6Attn: N. F. Rekos (RAP)

W. H. Roudebush (RL)Jack Suddreth (RL)

Department of the ArmyU.S. Army Aviation Material LaboratoryPropulsion Division (SAUFE-PP)Fort Eustis, Virginia 23602Attn: E. T. Johnson

Robert G. Dodd SAVDL-EU-PP , "

-1

10

11

111

-11

United Aircraft of Canada, Ltd.S.O. BoxLonquenil, Quebec, Canada .Attn: Miss Mary Cullen

Air Force Office of Scientific Research1 00 Wilson BoulevardArlington, Virginia 22209Attn: SREP

ArthurD Little, Inc.

Department of the NavyBureau of Ships . . . • ; - • • • 'Washington, DVC. 20360 'Attn: G. L« Graves 1

NASA-Langley Research CenterLangley StationTechnical Library -Hampton, Virginia 23365 'Attn: J . V . Becker - 1

United States Air Force .' • :

Aero Propulsion LaboratoryArea B, Bldg. 18DWright-Patterson A.I*.B'.Dayton, Ohio 5 33 • 'Attn: R. E. Henderson, AFAPL/TBC 1

Lt. William S. Blazowski, AFAPL/SFF 1

NASA-J. P. Kennedy Space Control (KSC)KSC, Florida 32899Attn: Mr. William Henry Lee

Mail Code LA-PLN 1

U« S. Navy Naval Air SystemsHdq., Washington, D.O, 20360 ,Attn: Mr. Robert Heitkotter Air 330D ; . 1

Naval Air Propulsidn Test CenterikkO Parkway AvenueTrenton, NJ 08628Attn: Mr. Anthony F.KLarman

Code PE-71 Building 3^ 1

Naval Air Rework FacilityNorth Island? Naval Air StationSan Diego, CA 92135Attn: Mr. Greg Marshal

Environmental Protection TeamCode 02210, Building 3 . ., • . . 1

Department of Transportation (DOT).4-00 7th Street, S.W.Washington, D.C.20590 :

Attn: Mr. Alan Grobecker CIAP 1Dr. Robert Underwood Code TST - 1

U. S. Air ForceLt. Larry LindenHdtr. USAF (PREV)Washington, D.O". 20330 !


Arnold Engineering & Development CenterAttn: Mr. E. L. Hively, Code DYRAF Arnold Station, TW 37389

Mr. Donald W. MaleAF Arnold Station, TN 37389

Kirtland Air Force BaseNew Mexico 87117Attn: Capt. Dennis F. NaugleAFWL (DEE)

U. S. ArmyDon Weidhuner .Army Material CommandWashington, D.C. 20315Attn: AMORD-GP

U. S. ArmySt. Louis, MO 63166Mr. L. E. BellAVSCOM, AMSAV-EFPBox 209

Environmental Protection AgencyMr. George D. KittredgehOl M.'Street, S¥Washington, D.O.

Dr. Lou. Schoen •401 M. Street, SWWashington, D.O.

Dr. Robert E. Sampson2565 Plymouth RoadAnn Arbor, MI kQl05

' Mr. J. R. SmithWERC • '.:Durham, HC 27701

Mr. John R. ZimmermanDIVMETResearch Triangle ParkRaleigh, NC 27711

Mr, John MoranOffice of Research and MonitoringResearch Triangle ParkRaleigh;, NC 277-11


Mr. John H. Wasser . •Environmental Protection AgencyResearch Triangle Park. NC -27711 :Wing "G"

National Aeronautics & Space Administration (NASA)Ames Research CenterMoffett Field, CA 9 35Attn: Mr. Weil Parlow

Mail Stop 2 5-5

Dr. I. G. PoppoffMail Stop 25 -1

NASA-Flight Research Center (FRC) P.O. Box 273Edwards, CA 93523Attn: Mr. Jack Nugent

Code R

Jet Propulsion Laboratory (JPL)800 Oak Grove DrivePasadena, CA 9H03Attn: Dr. James King

Mail Stop 183-toL

DOT/Transportation Systems CenterAttns Mr. Anthony Broderick (Tl)55 BroadwayCambridge, MA 021 2

Bureau of MinesBartlesville Energy Research CenterP.O. Box 1398Bartlesville, OK 7^003Attn: Mr. Richard Hurn

.United Aircraft Corporation. ,; . . . . •Pratt & Whitney Aircraft Division1+00 Main StreetEast Hartford,. Connecticut 06108Attn: Mr. R. Marshall

United Aircraft ResearchEast .Hartford, ConnecticutAttn: Library .


Detroit Diesel Allison DivisionDepartment 889 , Plant 8 .P.O. Box 89kIndianapolis,:Indiana 6206Attn: Mr. Prank Verkamp 1

Northern Research & Engineering Corp.219 Vassar Street .Cambridge, Massachusetts 02139Attn: K. Ginwala 1

General Electric CompanyFlight Propulsion DivisionCincinnati, Ohio 5215Attn: Technical Information Center N-32 . 1

. D. Bahr 1

General Electric Company :1000 Western Avenue 'West-Lynn, Massachusetts 01905Attn: Dr. C. W. Smith

Library Building . . . . ' , .

Curtiss-Wright CorporationWright Aeronautical DivisionWood-Ridge, NJ 07075 .Attn: D. Wagner . 1

AiResearch Manufacturing Company9851 Sepulveda Boulevard .Los Angeles, California 90009Attns L. C. Wright . 1

AVCO Corporation .Lycoming Division .550 South Main Street ; ..Stratford, ConnecticutAttn: C. W. Bolton 1

Continental Aviation & Engineering Corporation12700 Kercheval . . .Detroit, Michigan 1+8215Attn: E. H. Bernstein 1

International Harvester Company . . .Solar Division2200 Pacific HighwaySan Diego, California 92112Attn: P. A. Pitt 1

Arthur b Little, Inc.

George Derderian - AIR 536626Naval Air Systems Command l • 'Department of the NavyArlington, Virginia 20360 1

The Boeing Company ' . ' ' ' . "Commercial Airplane DivisionPloO. Box 3991Seattle, Washington 98124Attn: G. E. Hays 1

The Boeing CompanyMissile and Information Systems Division224 N. Wilkinson StreetDayton, Ohio 45402Attn: W. K. Thorson . : 1

Aerojet-General CorporationP.O. Box 1947.Sacramento, California 95809Attn: M. S. Wylin . 1

Cornell Aeronautical Laboratory :4455 Genessee StreetBuffalo, New York 14221 . 1

Marquardt Corporation16555 Saticoy StreetVan Nuys, California 91404 1

Thompson Ramo Wooldridge23555 Euclid AvenueCleveland, Ohio 1

Cummings Engine Company •1000 Fifth AvenueColumbus, Indiana 47201Attn: Roger Bascom 1

Garrett/AlResearch Company402 South 36th StreetPhoenix, Arizona 85034Attn: J. M. Hassis 1

Pratt & Whitney AircraftFlorida Research & Development CenterBox 2691 •West Palm Beach, Florida 33402Attn: G. Lewis 1


Aerojet General CorporationSacramento Facility -P.O. Box 158V7Sacramento, California 95813Attn: Dave Kors 1

The University of Toledo : .•.Toledo, Ohio 1*3606 •Attn: Dr. Duen-Ten Jeng 1 -

Eaton Yale andTowne Research Center26201 Northwestern Highway . .Southfield., Michigan 1*8075 .; • . •Attn: H. M. Reigner 1 '

RocketdyneNorth American Rockwell ;6630 Canoga Avenue ; .Canoga Park, California 9130!* .Attn: S. D. Clapp 1

Manager, Propulsion Technology Research Division

Department of Mechanical Engineering .201 Engineering Building 'Michigan State UniversityE. Lansing, Michigan 1*8823 • . "•Attn: Dr. M. C. Potter . '1

Clemson UniversityClemson, South CarolinaAttn: Dr. Tah-teh Yang . . 1 .

Cook Engineering Bldg.

John A. Belding - ,ATR 330P •. ,Naval Air Systems CommandDepartment of the NavyArlington, Virginia 20360 ' .!'•••

McDonnell-Douglas Company - .•Lambert-St. Louis Municipal AirportP.O. Box 516 . ' .St. Louis, Missouri .63166' .. ••.Attn: P. D. McVey • ' • 1

Ling-Temco VoughtMissiles & Space DivisionP.O. Box 6267Dallas,Texas 75222Attn: R. 0. Guthrie 1


Parker-Hannifin Company1T325 Euclid AvenueCleveland, Ohio Mj-112Attn;: Harold C. Simmons 1

The Johns Hopkins UniversityApplied Physics Laboratory8621 Georgia AvenueSilver Springs, Maryland 20910Attn: G. Dugger 1

Stanford Research Institute19722 Jamboree BoulevardIrvine, California 9266*1Attn: Dr. D. C. Wooten 1

Stanford Research InstituteMenlo Park, California 9*4025Attn: F. E. Weil 1

Delavan Manufacturing Company811 - Fourth StreetWest Des Moines, Iowa 50265Attn: R. J. Slezak 1

Aerospace Corporation2350 East El Segundo BoulevardEl Segundo, CaliforniaAttn: Owen W. Dykema 1

Department of Mechanical EngineeringMITCambridge, Massachusetts 02139Attn: Professor J. Heywood 1

University of CaliforniaDepartment of Mechanical Engineering • .Berkeley, California 9*4-720 'Attn: Professor R. F. Sawyer . ' 1

Northwestern UniversityDepartment of Mechanical Engineering & Astronautical SciencesEvanston, Illinois 60201 .Attn: Professor A. A. Kovitz 1


General Applied Science LabsMerrick and Stewart AvenuesWestbury, New York 11590Attn: Dr. R. Edelman ' 1

Coordinating Research Council, Inc.Thirty Rockefeller PlazaNew York, New York 10020 'Attn: Timothy C. Belian . 1

Gulf Research and Development CenterP.O. Drawer 2038Pittsburgh, PA 15230

• Attn: Dr. R. S. Spindt . 1

General Motors Corp Research LaboratoriesFuels and Lubricants Depto12 Mile and Mound RoadsWarren, Michigan U8090 -Attn: Dr. G. J» Barnes ' 1

Mobil Research and Development CorporationResearch Department - • . •Paulsboro, New Jersey 08066 . :

Attn: F. J. Hills 1

Energy Research Corporation • '15 Durant Avenue • . .Bethel, Connecticut 06801 /Attn: Richard E. Engdahl • 1

Environmental Protection AgencyNERC Room 0 38Research Triangle Park, North Carolina 27711Attn: John H. Sigsby 1

J. M. PerezResearch Department, Technical CenterCaterpillar'Tractor Co.Peoria, Illinois 6l602 • 1


NASA Lewis Research Center •21000 Brookpark RoadCleveland, OH 44135Attn: Contract Section A/MS 500-206 ' 1

Patent Counsel/MS 500-113 1Lewis Library/MS 60-3 2Report Control Office/MS 5-5 1W. L. Stewart/MS 501-5 1J. B. Esgar/MS 60-4 1W. T. Olson/MS 3-16 1R. A. Rudey/MS 60-6 1J. F. Dugan/MS 86-1 1S. Lieblein/MS 501-5 . .1.R. E. Jones/MS 60-6 1J. Grobman/MS 60-6 1Lt. Col. G. S. Weden/MS 500-317. 1Louis Rosenblum/MS 302-1 . 1.J. Stuart Fordyce/MS 309-1 1Dr. John J. Gulan/MS 5-4 1Edward F. Stevenson/MS 21-6 1Robert Branstetter/MS 60-6 1Robert Manly/MS 21-6 1Erline Trsek/MS 302-1 1Patricia O'Donneil/MS 309-1 1Evelyn Anagnostou/MS 302-1 1Robert Summers/MS 7.7-1 1John Toma/MS 302-1 1Andrew Leavitt/MS 500-120 1Al Evans/MS 6-1 1H. F. Butze/MS 60-6 25


prepared for - nasa · 2013-08-31 · h 7 3 2 3 0 9 8 nasa cr-121159 adl 74443 odor intensity and...

Documents