Soot formation in laminar diffusion flames at elevated temperatures

SOOT FORMATION AT HIGH TEMPERATURES effects, the measured flame temperature of the diluted flame is brought back to the undiluted fuel flame temperature by replacing a portion of the nitrogen in air with argon. Although this method provided data identifying the separate effects of temperature and dilution, and the results showed that the relative importance of dilution and temperature varies with the amount of diluent addition, there remains one concern: The temperature field of a diluted flame may not be representative of an undiluted flame whose temperature is altered by changing the temperature of the reactants. This point was noted by Boedeker and Dobbs [4], who observed an inversion in the axial temperature profiles where the nitrogen diluted flames are cooler near the burner but hotter at the top, when compared with undiluted flames. However, this effect was believed to be, at least in part, due to particulate radiative loss a consequence of the change in the amount of soot [4]. In order to avoid the overlapping effects of dilution and temperature, Giilder and Snelling [8] opted for an experimental arrangement in which the co-annular air and fuel streams are heated to elevate the flame temperature. They report the influence of flame temperature on soot information in propane and isooctane flames burning at their smoke point heights. In the present study, which complements the previous work [8], we focus on (a) the influence of flame temperature on soot formation in flames in which the fuel flow rate was kept constant as the flame temperature was varied, and (b) the temperature sensitivity of different fuels. Also reported are the data obtained on flames burning at their smoke points at all temperatures considered. Propylene, ethylene, and isooctane were used as fuels. Data obtained with propane in our previous study [8] have also been used in the assessment of the temperature sensitivity of different fuels. EXPERIMENTAL METHODOLOGY The fuel nozzle of the burner is a stainless-steel pipe of 3 mm inner diameter. Air is supplied from a concentric converging nozzle of 100 mm inner diameter. Both air and fuel streams are heated by regulated electric heaters. Temperatures of the reactants are monitored by thermocouples near the exit of the fuel and air nozzles, and kept within + 4 K of the desired reactant

75 temperature. The fuel flow rate is monitored by calibrated rotameters. The air, before exiting from the converging nozzle, passes through a bed of glass beads and a set of wire-mesh screens to prevent flame instabilities. A flame enclosure with glass windows provides optical access and protects the flame from air movements in the room. The burner assembly sits on a positioning platform with accurate vertical and horizontal movement capability. The line-of-sight average soot volume fractions along the centerline of the flames were measured by the transmission r(k) of a multiline laser beam made of three wavelengths. GaA1As (830 nm), H e - N e (632.8 nm), and A r - I o n (515 nm). The optical path length L that is, the flame diameter, was measured by a reading telescope with an eyepiece. Typical flame diameters at the lower half of the flames were around 6 - 8 mm. Each division of the scale on the eyepiece corresponds to 0.125 mm. The repeatability and the reproducability of the flame diameter measurements were within + 1 scale division with gaseous fuel flames and + 2 divisions with isooctane flames. The soot volume fractions can then be calculated assuming Rayleigh extinction as follows X/L F

=

Im{(m z-

1)/(m z+2)}.67r

× In r ( k ) ,

(1)

where k is the wavelength. The complex refractive index of the soot particles was taken as m = 1.89 - 0.48i [9], to have data based on the same assumptions as in our previous soot work [10]. The spectral radiance of a blackbody is given by the well-known Planck radiation law, which can be simplified, for the wavelengths and temperatures of interest in this work, to the Wien radiation law ¢1

NO(k, T) = kS{exp(c 2 / X T )

-

C1

k5 exp(c 2 / k T )

1} (2)

'

where N ° ( k , T) is the spectral radiance, k is wavelength, T is the blackbody temperature, and

76

O . L . GULDER

c~ and c z are constants: c~ = 1.909. 10 -2 W cm2/sr, and c 2 = 1.4388 cm K. For an emitting body with a spectral emissivity e(h, T), the spectral radiance or brightness temperature To at a wavelength X, which is defined as the temperature of a blackbody that has the same spectral radiance as the emitter is given by

1/Ts=

(I/T)

- ( X / C z ) . In ~(X, T ) .

(3)

The spectral emissivity of a distributed source of radiators is given by [11] e()x) = 1 - r ( X ) ,

(4)

where r(h) is the spectral transmissivity over the specified optical path. If both the soot brightness temperature and the flame transmission are measured at a particular wavelength and over the same path length and if the temperature is constant, then Eqs. 3 and 4 can be used to obtain the soot surface temperature T. If the temperature is not constant over the optical path length then the measured temperatures are average values weighted by soot concentration and should be representative of those in soot-laden regions of the flame. The instrument used to measure the soot brightness temperature is a Micro-Optical Model No. 95 disappearing filament pyrometer. This instrument was modified to be used for both red (655 nm) and green (549 nm) brightness temperature measurements. The pyrometer was calibrated using a model GE 18A/T10/1-6V strip filament lamp whose brightness temperature at 654 nm was measured to an accuracy of < 5 K as a function of lamp current. The total temperature uncertainty of the pyrometer calibration is +_15K. To minimize the effect of uncertainties r(X) on the determination of T, the best-fit curves to soot volume fraction data were used to obtain flame transmissivities from Eq. 1 for isooctane flames. For other flames, transmissivities at 549 and 655 nm were obtained from the measured transmissivities at 830, 632.8, and 515 nm by interpolation/extrapolation. The soot temperature was then calculated using Eqs. 3 and 4 at both green and red wavelengths giving two independent estimates of the temperature. The temperatures obtained from the brightness temperatures and

emissivities at 549 and 654 nm were in good agreement (maximum difference less than 20 K), and they displayed the same trend. Reported in this article are the temperatures obtained at 654 nm. Further details of the experimental rig and the soot temperature measurements are given in Ref. 8. The line-of-sight average light extinction, the brightness temperature, and the optical path length were measured, as a function of the axial position, along the centerlines of propylene, ethylene, and isooctane diffusion flames. Experimental conditions are given in Table 1. Measurements were made at three reactant temperatures, namely 300, 473, and 623 K. Additional soot measurements at 680 K with propylene and at 673 K with ethylene were also conducted. In the first set of experiments, fuel flow rates for a given fuel were kept constant at all reactant temperatures. In the second set, the fuel flow rates were adjusted to TABLE 1 Experimental Conditions and Summary of Results a

Tr

mI

SP

Fma,

T~d (TAmax

Fuel

(K)

(mg/s)

(mm)

(ppm)

(K)

(K)

Ethylene

300 473 473 623 623 673 300 473 473 623 623 680 300 473 473 623 623 300 473 653

3.0 2.8 3.0 2.6 3.0 3.0 0.97 0.9 0.97 0.85 0.97 0.82 1.56 1.21 1.56 1.05 1.56 6. 4.8 3.8

70 64 N/Aa 56 N / A t' N/A b 26 22 N/A b 19 N/Aa 17 43 27 N/Ab 22 N/A ~ 145 92 63

7.8 9.6 10.1 10.3 12.1 13.3 13.7 16. 17.8 19.4 22.6 20.6 9.5 10.2 13.2 13.3 19. 5.4 6.2 6.9

2372 2441 2441 2500 2500 2520 2337 2409 2409 2471 2471 2515 2274 2352 2352 2418 2418 2269 2347 2427

1795 1825 1805 1865 1835 N/M c 1745 1745 1765 1760 1760

Propylene

Isooctane

Propane

N/M e

N/M c 1790 1770 1805 1805 1690 1760 1810

aPropane data are from Giilder and Snelling [8]. T r is the temperature of the reactants, m f the fuel flow rate, SP the smoke point height, Fma× the line-of-sight maximum soot volume fraction, Tad adiabatic equilbrium flame temperature, and (Ts)max the line-of-sight maximum soot surface temperature (maximum surface temperature reached before midheight of the flame). bNot applicable. CNot measured.

SOOT FORMATION AT HIGH TEMPERATURES obtain s m o k e p o i n t temperatures.

flames at all reactant

RESULTS AND DISCUSSION Constant Fuel Flow Rate Conditions

The line-of-sight average soot volume fraction profiles, as a function of axial position, of ethylene flames at four different reactant temperatures are shown in Fig. 1. The fuel mass flow rate is 3 mg/s in all cases. This rate corresponds to a smoke point flame at T, = 300 K, but at higher reactant temperatures some amount of unoxidized soot escapes from the tip of the flame (Fig. 1). The change in the maximum soot volume fraction for a change in reactant temperature from 673 to 300 K is about 40%, and the corresponding change in adiabatic flame temperature is about 148 K (Table 1). Santoro and Semerjian [3] studied the effect of flame temperature on soot formation in a concentric co-flowing diffusion flame by adding nitrogen to the fuel and keeping the fuel flow rate constant. They added sufficient nitrogen (nitrogento-fuel ratio 2.29 by volume) to ethylene to reduce the adiabatic flame temperature from 2369 K (for pure ethylene) to 2223 K. A depression of 146 K in the adiabatic flame temperature of this ethylene flame provided an order of magnitude reduction, from 11 ppm to approximately 1 ppm, in the

E D . 20

77 amount of maximum soot concentration. In our experiments with ethylene, at constant fuel flow rate, the change in the adiabatic flame temperature from Tr = 673 K to T, = 300 K is 148 K (Table 1). Corresponding change in the maximum soot volume fraction is very small as compared with the results of Santoro and Semerjian [3] for the same amount of temperature change. The maximum soot volume fraction reduces to 7.8 from 13.3 ppm, a decrease about 40% (Fig. 1 and Table 1). Since the two studies were carried out on similar diffusion flames, the only possible explanation for this significant discrepancy is the dominant influence of dilution on soot formation due to the inert diluent added to the fuel in the study of Santoro and Semerjian [3]. It should be noted that in the present work and in Ref. 3 temperatures cover different domains. Then, the foregoing comparison is semiquantitative due to the nonlinearity of the effect of temperature on soot formation. The soot volume fraction variation in propylene flames with temperature is quite similar to that of ethylene flames (Fig. 2). A fuel flow rate of 0.97 mg/s at T, = 300 K yields a smoke point flame, and at higher reactant temperatures soot escapes from the tip of the flame. Soot particle surface temperatures measured in ethylene and propylene flames at different reactant temperatures are shown in Figs. 3 and 4, respectively. It should be noted here again that

E 30

Q.

Q.

=._o 15

I Ethylene, 3 mgls

.-D-T-a00 K --~-- T - 4 7 3 K .-m- T - 6 2 3 K --~- T-673 K

o

.o

25

Propylene

0.97 mg/s

--c'- T - 3 0 0 K .-~-- T - 4 7 3 K ~

T-flgR

K

2o

P U.

® 10 E

m 15 E ~ 10 >

O

>

5

iriT o o co

O o

0

5 0

0

10

20

30

40

50

60

70

80

5 10 15 20 25 30 Height Above the Burner, mm

Fig. 1. The line-of-sight average soot volume fraction profiles, as a function of axial position, of ethylene flames at four different reactant temperatures. Fuel flow rate was kept constant, at 3 mg/s, for all temperatures.

Fig. 2. The line-of-sight average soot volume fraction profiles, as a function of axial position, of propylene flames at three different reactant temperatures. Fuel flow rate was kept constant, at 0.97 mg/s, for all temperatures.

Height A b o v e the Burner, mm

78

O.L. GOLDER

these are line-of-sight temperatures that represent averages over the flame diameter. Also shown in Fig. 3 are the soot particle surface temperatures reported by Flower [12], who employed a similar method for determining soot temperatures in a flame with ethylene flow rate of 3 m g / s at 0.1 MPa and room temperature. The temperature profile measured by Flower [12] and the profile obtained in the present work for T r = 300 K are quite similar except that there is an axial shift of about 10-12 mm between the two (Fig. 3). The major cause of this axial shift is the different diameters of the burners used in these two studies. In Flower's experiment, the inner diameter o f the burner was 12.7 mm, whereas in the present study it is 3 r a m . As a result of this the visible flame height was about 56 mm in Flower's experiment (Fig. 3a of Ref. 12), whereas it is 70 mm in the present experiments (Fig. 1). The temperature profiles are axially compressed/ expanded in accordance with their respective visible flame heights, although the self-similarity of the temperature fields was preserved. The maxima and the minima of the two profiles are in close agreement (within 15 K). In lower portions of the ethylene flames, soot surface temperatures are higher at higher reactant temperatures. The maximum soot temperature location shifts towards the burner as the reactant temperature is increased. Near the midheight of the ethylene flames, the temperatures are almost the same and they remain mostly identical from midheight to the flame tip (Fig. 3). A higher radiative hee.t loss

1900 1850

Propylene 097 mgls I''D-T ~-m-T! ---o~l" " 473300823 KK/K

1800 P 1750 E1700 I--

1650

~1800

1550 1500

~ i r T' ~ , r ,I 'r~r[~

5

10

f~l

15

IT rrT[~

20

r F~ l

25

30

Height Above the Burner mm Fig. 4. The line-of-sight average soot surface temperature profiles, as a function of axial position, of propylene flames at three different reactant temperatures. Fuel flow rate was kept constant at 0.97 mg/s. from high reactant temperature flames is the major cause of this behavior. The variation of the soot temperatures with axial position and the reactant temperature is quite different in propylene flames (Fig. 4). Soot temperatures at Tr = 473 and 623 K are lower than those at Tr = 300 K in the upper portion of the flames. Further, at all axial positions, soot temperatures at T r = 623 K are lower than those at T~ = 473 K (Fig. 4). This is expected because of the high soot loading of the propylene flames, especially at higher temperatures, as compared with soot concentrations in ethylene flames (Figs. 1 and 2). If the maximum soot volume fraction

1850 b~

1800

E @

" 1 ~ ~

~

1750

I.Q

o a 1700 ,,,,,j ¢/J o

1650

- ~ - T = 300 - ~ - T - 473 - ; ~ - T - 623 " Flower'.

K K K Data

-mmn~

Ethylene 3 mg/s ,,~,/,,,,),,,,/,,,,f,,,,t,,,ll,,,,

0

10

20

30

40

50

Height Above the Burner mm

60

70

Fig. 3. The line-of-sightaverage soot surface temperature profiles, as a function of axial position, of ethylene flames at three different reactant temperatures. Fuel flow rate was kept constant at 3 mg/s. Black square symbols are the data from Flower [12] measured at room temperature on a burner with 12.7 m m inner diameter.

SOOT FORMATION AT HIGH TEMPERATURES per unit mass of fuel flow is taken as a characteristic soot concentration, this parameter is approximately six times larger for propylene flames than for ethylene. However, low in the flame, gas temperatures are expected to follow the ordering of the adiabatic flame temperatures corresponding to different reactant conditions. Our limited thermocouple measurements in propane flames with different reactant temperatures, and centerline temperature measurements of Gomez et al. [13] in diluted flames support the assumption of qualitative agreement of the true gas temperatures in the lower regions of the flames with adiabatic flame temperatures. The behavior of the soot temperature profiles in the upper portions of the flames is the result of the competition between heat generation upon soot oxidation and heat loss due to radiation. Around midheight of the flames, temperatures start decreasing with height due to still high soot concentrations. At higher axial positions soot concentrations are lower, resulting in lower heat losses by radiation, and if the gas temperatures are high enough to sustain soot oxidation then soot surface temperatures start increasing near the tip of the flame (Figs. 3 and 4). Smoke Point Conditions

Variation of the soot volume fraction, with axial position and the reactant temperature, in ethylene flames burning at their smoke point heights is shown in Fig. 5. To maintain the flame at its smoke point as the temperature is elevated the fuel flow rate must be reduced. The observed flow rates are noted in Fig. 5. The maximum soot volume fractions at elevated temperatures under smoke point conditions are lower (Fig. 5) than those at constant fuel flow rate conditions (Fig. 1). However, the maximum soot volume fractions per unit of fuel supplied to the flame are almost identical to those at constant flow rate conditions. Propylene and isooctane flames at elevated temperatures exhibit behavior similar to ethylene flames (Figs. 6 and 7). Gomez et al. [131 examined the influence of flame temperature on soot formation in smoke point butene diffusion flames by changing the flame temperature through nitrogen addition to the fuel. Since propylene and isobutene have similar sooting characteristics

79

E =12.

3

1, itEthylene 12

|Smoke Point Condition== 1

•-E-- T-300 K --~-- T-473 K - - l - T-623 K

a

I.I.

•E

6

"~ >

4

E

2 0

0

10

20

30

40

50

60

70

80

Height Above the Burner, mm Fig. 5. Variation of the soot volume fraction, with axial position and the reactant temperature, in smoke point ethylene flames. Numbers on the curves are the fuel mass flow rates at corresponding reactant temperatures.

[14], present results obtained on propylene flames can, at least semiquantitatively, be compared with the data of Gomez et al. [13]. Since both sets of data were obtained with smoke point flames (by adjusting the fuel flow rate for all conditions in both studies) and the fuel flow rates are not the same, the comparison will be based on the maximum soot volume fraction per unit fuel flow to the burner, Fma x/mf (ppm/mg/s). They added nitrogen as an inert diluent to the fuel to reduce the adiabatic flame temperature from 2318 to

30 ,,

Propylene

¢~ 25

F ~-o-T-300 K ~T-473

Ki

SmoRe Point Conditions :--B,-T-623 K

0

i --~-- T - 6 8 0 K

° 20 m

10.83 rng/$~,~,~,~.

J

"~ 10 0

0

-

0

5 10 15 20 25 30 Height A b o v e the Burner, mm

Fig. 6. Variation o f the soot volume fraction, with axial position and the reactant temperature, in smoke point propylene flames. Numbers on the curves are the fuel mass flow rates at corresponding reactant temperatures.

80

O.L. GULDER 20

.

.

.

.

,

.

.

.

.

,

.

.

.

.

,

.

.

.

.

,

.

.

.

.

f

.

.

.

.

,

.

.

.

.

isooctane O.

Smoke Point Conditions

P = 1o

LL

1.05

• •

T -- 473 K T = 623 K

mg/s~

--I

"8

g,

5

0 0

5

10

15

20

25

30

35

Height Above the Burner, mm 2133 K with a nitrogen-to-isobutene ratio of 5.7. The adiabatic flame temperature is depressed, by the addition of nitrogen, by 185 K. This 185 K change in temperature is accompanied by a substantial decrease in maximum soot volume fraction per unit fuel flow, from 6.8 to 1.2 ppm/mg/s (Table 1 and Fig. 6 of Ref. 13). In the present experiments with propylene, the maximum increase in adiabatic flame temperature is about 178 K (Table 1), but the increase in maximum soot volume fraction per unit fuel flow due to this temperature change is less than two fold, from 14.1 to 25.1 ppm/mg/s (Table 1). The discrepancy between the two results is due to the dilution effect accompanying the temperature effect as a result of the nitrogen added to the fuel in the study of Gomez et al. [13]. It should be noted again that this comparison is semiquantitative due to nonlinearity of the temperature effect. Soot particle temperature profiles of smoke point ethylene, propylene, and isooctane flames at different reactant temperatures are shown in Figs. 8-10, respectively. Soot surface temperatures in the lower half of the flames are similar to those at constant fuel flow rate conditions for ethylene (Figures 3 and 8) and for propylene flames (Figures 4 and 9). In the upper halves of the ethylene and propylene flames, soot temperatures are generally higher at higher reactant temperatures, (Figs. 8 and 9). Comparison of the soot temperature profiles of propylene at T~ = 623 K (Figs. 4 and 9) points to the significant influence of the fuel flow rate on soot temperatures. The

Fig. 7. Variation of the soot volume fraction, with axial position and the reactant temperature, in smoke point isooctane flames. Numbers on the curves are the fuel mass flow rates at corresponding reactant temperatures.

behavior of the soot temperature profiles in isooctane flames is very similar to that in propylene flames (Fig. 10).

Temperature SensitivityofFuels In order to illustrate the response of the sooting behavior of different hydrocarbons to flame temperature changes, normalized maximum soot volume fraction values are plotted against increases in adiabatic flame temperatures in Fig. 11. Normalized maximum soot volume fraction is the ratio of the maximum soot volume fraction per unit fuel flow rate (ppm/mg/s) at a given reactant temperature normalized with respect to the value 2000 v

1950

Ethylene Smoke point conditions

473 623

--~-- T

1900

i

i

1850 2.8 regis

1800

~- 1750 ¢n 1700 1650 1800

- ~ ' 1

0

....

i , , , , l , , , r ~ , , , i , , , , i

10 20 30 40 50 60 Height Above the Burner, mm

....

70

Fig. 8. Variation of the average soot surface temperature of the smoke point ethylene as a function of axial position and the reactant temperature.

SOOT FORMATION AT HIGH TEMPERATURES

1900 1850

2000

1

___jo.e-~

v

mgls

i

•

1900

:

1550

--e-T

T ~ 0

....

r, ~

~

.

5 10 15 20 25 Height A b o v e the Burner, mm

30

Jr~

my

(5)

The adiabatic flame temperature increase is the difference between the adiabatic flame temperature at a given reactant temperature and the adiabatic flame temperature at Tr = 300 K. These values were calculated using the data given in Table 1, and the results are shown in Fig. 11. 2.2

i

2.0

ii

A

Ethylene_SP

• 1~ •

Ethylene Propylene_SP

• (3

o

Or)

i

./11 •

/ /

e"

160O

Isooctane Smoke Point Conditions

i

I

i

5

i

i

i

,

i

,

L

10 15 20 25 30 Height Above the Burner, mm

,

35

i

V// A l k a n e s © /

Propylene Isooctane_SP Isooctane Propane_SP

V

•"O 1.8

i

.

Data obtained with the three fuels used in this study, and the propane data from our previous study [8], indicate that olefinic fuels (ethylene and propylene) exhibit lower temperature sensitivity than alkanes (propane and isooctane) (Fig. 11). Gomez and Glassman [2] noted that aromatics and dienes show lower temperature sensitivity than the aliphatics, but they did not differentiate among different aliphatic groups. To our knowledge, this is the first observation of the differing temperature sensitivities of the two aliphatic groups. However, one cautionary remark is in order. Ethylene and propylene have flame temperatures 60-100 K higher than propane and isooctane for the same initial reactant tempera-

Nf ]Tr=300K

roe n

.

Fig. 10. Variation of the average soot surface temperature of the smoke point isooctane flames as a function of axial position and the reactant temperature.

Imaxl t

.

ilimlaiio_ea~•,.~ ~ • . t -

~, , , ~7-~

= 300 K, that is,

/~max =

/ /

/

E 1.6 x

.

623 K

Fig. 9. Variation of the average soot surface temperature of the smoke point propylene flames as a function of axial position and the reactant temperature.

._.

.

T = 623 K

1800

1 500

o

.

T = 473 K

•

O O

(/'3

Tr

.

E I~ 1700

3

at

.

J 0) O.

0.97 regis

1500

.

0.9 mgls

o.

~" 1650

.

Propylene, smoke point oondltlon=

1000 ~ =l

81

/ [] &." . " Olefins

1.4

"D

•

1.2

E

1.0

,,,,

0.8 -50

0

/~'./

O

Z

I

I

I

50

I

I

I

100

Adiabatic Flame Temperature

I

150

I

I

200

Increase, K

250

Fig. 11. Temperature sensitivity of the alkanic and olefinic hydrocarbons studied in this work. In the legend, fuel names followed by an SP refer to a smoke point flame. Others are flames at fixed fuel flow rates at different reactant temperatures. Normalized maximum soot volume fraction is defined in the text. Adiabatic flame temperatures corresponding to various reactant temperatures are given in Table 1.

82 tures. The increase in soot formation rate in diffusion flames with temperature levels off at around 2000-2200 K and then starts decreasing, forming a bell-shaped curve [15]. Therefore, a small portion of the difference in sensitivities shown in Fig. 11 could be attributed to the relatively higher flame temperatures of ethylene and propylene. The pyrolysis rates and intermediates are expected to be different due to different chemical structures of olefins and alkanes. These also would contribute to the difference observed in temperature sensitivities.

6. L. GULDER

o f the data o f Ref. 13 in an earlier version o f this article. I am indebted to Dr. D. R. Snelling f o r implementing the soot surface temperature measurement system, and advice on optical matters. The capable assistance o f Mr. M. F. Baksh in data collection is gratefully acknowledged. The work described herein was supported by National Research Council's internal f u n d s (program manager Mr. L. Gardner) and by the Canadian Governments E M R / P E R D program (program manager Mr. P. ReillyRoe). I thank Mr. L. Gardner f o r his support and encouragement.

CONCLUSIONS

REFERENCES In this work, we measured the line-of-sight soot volume fractions and soot surface temperatures as a function of axial position in laminar diffusion flames. The measurements were made at various temperatures of the reactants to study the influence of flame temperature on soot formation and the temperature sensitivity of different fuels. Obtained data from smoke point flames and flames at fixed fuel flow rates at different reactant temperatures using ethylene, propylene, isooctane, and propane as fuels lead to the following conclusions: 1. Olefins (propylene and ethylene) show a lower sensitivity to flame temperature than alkanes (isooctane and propane). 2. Maximum soot volume fractions and soot formation rates increase with increased flame temperature, as noted by others, but the observed changes are not as large as the changes measured by other investigators by adding an inert diluent to the fuel to alter the flame temperature. In previous diffusion flame studies in which the temperature effects were examined by nitrogen dilution of the fuel, the observed effects on soot formation are due to dilution as well as the temperature. For this reason, results of such previous studies should be reevaluated. 3. In flames of fuels with high sooting propensities, the line-of-sight soot surface temperatures of a high Tr flame can be lower than those of a lower Tr flame, mainly due to intense radiative heat looses.

1 am grateful to an a n o n y m o u s reviewer who pointed to an error in m y interpretation

1.

2.

3.

4.

5. 6.

7.

8.

9.

10. 11.

12. 13. 14.

15.

Schalla, R. L., and McDonald, G. E., Fifth Symposium (International) and Combustion, Reinhold, New York, 1955, p. 316. Gomez, A., and Glassman, I., Twenty-First Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1987, p. 1087. Santoro, R. J., and Semerjian, H. G., Twentieth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1985, p. 997. Boedeker, L., and Dobbs, G. M., Twenty-First Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1987, p. 1097. Axelbaum, R. L., Flower, W. L., and Law, C. K., Combust. Sci. Technol. 61:51-73 (1988). Du, D. X., Axelbaum, R. L., and Law, C. K., Twenty-Second Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1991, p. 387. Axelbaum, R. L., and Law, C. K., Twenty-Third Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1991, p. 1517. Giilder, t~. L., and Snelling, D. R., Twenty-Third Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1991, p. 1509. Lee, S. C., and Tien, C. L., Eighteenth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1981, p. 1159. Giilder, t~. L., Combust. Flame 78:179-194 (1989). Penner, S. S., Quantitative Molecular Spectroscopy and Gas Emissivity, Addison-Wesley, Reading, MA, 1959, p. 179. Flower, W. L., Combust. Flame 77:279-293 (1989). Gomez, A., Littman, M. G., and Glassman, I., Combust. Flame 70:225-241 (1987). Schug, K. P., Manheimer-Timnat, Y., Yaccarino, P., and Glassman, I., Combust. Sci. Technol. 22:235-250 (1980). Frenklach, M., Clary, D., Gardiner, W. C., and Stein, S. E., Twentieth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1985, p. 887.

Received 15 May 1991; revised 3 September 1991