Radiation from Luminous Flames'

---

INDUSTRIAL A N D ENGINEERING CHEMISTRY

4

Vol. 19, No. 1

Radiation from Luminous Flames’ By R. T.H a s l a m and M. W. Boyer DEPARTMENT OF CXEMICAL ENGINEERING, MASSACHUSETTS INSTITUTE OF TECHXOLOGY, CAMBRIDGE, MASS

T

HE problem of heat transfer is attaining a position of ever increasing importance to the chemical engineer. Of the three methods by which heat can be transferrednamely, conduction, convection, and radiation-the first two have been given the greatest attention from the point of view of practical application. Equations have been derived or set up empirically, which permit calculations giving a fair degree of accuracy. In the field of radiation there is the Stefan-Boltzman law for the calculation of radiation between solids. The use of this law involves a knowledge of the per cent “blackness” of the body, about which relatively little is known. The lack of accurate knowledge concerning radiation from gases and flames is even greater. Flames are classed either as luminous or non-luminous. Luminosity depends upon the amount of deposited solid matter in the flame, usually carbon, which is heated to incandescence, giving the flame its brightness. I n the nonluminous flame the finely divided solid matter is absent. While it has been generally known that a luminous flame radiates more heat than a non-luminous flame, very little experimental work has been carried out to determine the quantitative relation between the amounts of heat radiated from the two kinds of flames. It is the purpose of this paper to give a brief survey of the present knowledge concerning radiation from flames, particularly luminous flames, to present the results of .experiments carried out in the Department of Chemical Engineering a t the Massachusetts Institute of Technology, in which radiation from various types of flames was measured, and to indicate briefly the significance of these results from an industrial point of view. Review of L i t e r a t u r e

In 1884 Fredrick Siemens designed a regenerative glass furnace in which the combustible materials were burned in a separate chamber from the charge and the products of combustion led into the furnace. Since the charge was out of contact with the flame, it was heated chiefly by radiation. This is believed to be the first time that radiation was recognized as an important factor in heat transfer from flames. Siemens was of the opinion that the radiation was due to incandescent carbon particles in the flame, and that there was little radiation from a non-luminous flame. Six years later von Helmholtz’#* attempted the accurate measurement of radiation from flames produced by burning various gases. Table I shows the results of his work. Table I HEATRADIATED GAS

Per cent Hvdroiren d r b o i monoxide moI Methane Acetylene Illuminating gas . . Petrole Petroleum

NON-LUMINOUS

LUMINOUS

Per cent

Per cent

3.63 6.17 11.5 8.5 18.2

8.74

5.15 6.12 5.12

It will be noted that the greatest increase in radiation (about 120 per cent) occurring in the luminous flame over the 1 Received July 22, 1926. Presented before the Division of Gas and Fuel Chemistry a t the 72nd Meeting of the American Cheniical Society, Philadelphia, Pa., September 5 to 11, 1926. Numbers in text refer to bibliography a t end of article.

*

non-luminous flame was in the case of ethylene. This is worthy of note since ethylene was the least saturated gas used and one in which the most thermal decomposition with the production of free carbon would be expected. Julius, who worked with Helmholtz, made a very complete analysis of the radiation from different kinds of flames. He found almost all of the radiation to fall into two bands having wave lengths of 4.4 P and 2.8 P, respectively. In a pure hydrogen flame the 4.4 band almost completely disappeared, while in a carbon monoxide flame the 2.8 band disappeared. This was taken as proof that in the nonluminous flame the source of radiation was the carbon dioxide and water molecules. As a result of this work, Helmholtz advanced the theory that the radiation from a non-luminous flame depended upon the amount of carbon dioxide and water formed during combustion, and used this as a basis for calculation of the radiation from non-luminous flames. This theory has been discussed by one of the present authors.2 About the same time Pashen3believed radiation from flames to be a purely thermal phenomenon rather than chemical. He endeavored to prove this by experimentation, but his work failed to bear out his contention. While attempting to improve the efficiency of internal combustion engines. Callender4became interested in radiation as one of the causes of heat loss in the engines. He repeated some of Helmholtz’s experiments and found that a non-luminous coal gas flame 30 mm. in diameter may radiate as much as 15 per cent of the whole heat of combustion. In commenting on the low values of Helmholtz, Callender states: According to my experiments this low value is to be explained by the fact that he employed in these measurements small flames (6 mm. diameter by 60 mm. high), which were probably burning a t a comparatively low temperature, and which do as a matter of fact give a percentage of this order.

He carried out further experiments with coal gas, and obtained values of 15 to 20 per cent for luminous flames as compared with 10 per cent for non-luminous flames. Measurement of the radiation emitted in the course of an explosion and subsequent cooling of a gas and air mixture in a closed vessel, by Hopkinson,4 showed the radiation to amount to over 22 per cent of the whole heat of combustion with a 15 per cent mixture of coal gas and air. Considerable work on radiation from flames has been carried out in the Department of Chemical Engineering a t the Massachusetts Institute of Technology. In 1923Hunneman6 investigated the effect of depth of flame upon radiation. From his results he was able to set up an exponential law expressing the relation between flame depth and radiation. He also found that preheating the primary air decreased the amount of radiation. The following year Lovel16 took up the problem of total radiation from flames, dealing with non-luminous flames only. He summarizes his work as follows: The total radiation decreases in all cases as more than the theoretically required air is premixed with the gas. The total radiation also decreases when insufficient air for complete combustion is used. The total radiation varies with different gases from 10 per cent to 15 per cent of the latent heat of the gas, and bears no apparent relation to the temperature of the flame, the products of combustion formed, or the over-all chemical reaction.

January, 1927

INDUSTRIAL AND ENGINEERING CHEMISTRY

5

In 1924 there appeared a r6sum6 of the present status of The mechanism of cracking gaseous hydrocarbons by burnknowledge in the field of radiation from flames and gases.’ ing them with insufficient air is common knowledge. It is Of special interest in the field of radiation from luminous believed that the carbon particles thus formed act as carriers flames is the work of Lent and Thomas on measurements of or forwarders of heat, so to speak. These tiny bits of carbon, radiation from blast-furnace gas and generator gas flames in deposited in the flame by the cracking process, are instantly a Martin oven. They found the radiation from a flame of heated up by conduction from the hot gases of the flame to blast-furnace gases, to which benzene had been added as an state of incandescence, in which condition they radiate heat illuminating agent, but which was added in such a way as to the cooler surroundings, just the same as a solid body. not to increase the flame temperature, to be four times the While in the case of non-luminous flames experimental results seem to indicate that radiation is to a large extent independradiation from a non-luminous flame of the same kind. In a more recent article Lents states that measurement of ent of flame temperature, a consideration of the above statethe radiation from a sooty blast-furnace gas flame gives ment shows that this, obviously, is not true for that part values similar to those from a black body a t the same tem- of the radiation from the luminous flame, which is due to the perature. He also suggests the inoculation of non-luminous carbon. For if the radiation from the carbon particles is flames with an illuminating agent as a method of increasing the same as that from solid bodies, which there is every reason the rate of heat transfer from non-luminous flames used in to believe is the case, then the radiation from the carbon p a r t i c l e s will follow the industrial work. Stefan-Boltzman law and While the agreement bebe a function of the fourth t w e e n t h e results of the Radiation from luminous flames of the four gasespower of the absolute temvarious investigators is not city gas, methane, ethylene, and acetylene-was studied perature of the flame. The always close, the results are, by allowing the radiation from a small flame to fall on exception to this case would nevertheless, of t h e s a m e a thermopile at such a distance from the flame that it be a thin luminous flame, in order of m a g n i t u d e . I n could be considered as a point source. The radiation which the projected area of e v e r y case the radiation from luminous flames of city gas, methane, and ethyltne carbon particles would from the luminous flame is ene amounts to 9.2,8.0, and 10.65 per cent of the total not form a continuous area. considerably greater than heat of combustion, respectively. These values, howThe difficulties of accurately that from the non-luminous ever, are considered to be low, because of the difficulty measuring the true temperflame and bears out the imof obtaining complete luminosity in such a small flame. ature of a flame have, so portance of incipient carbon Acetylene gave a truly luminous flame, the radiation far, prevented the verificaformation in heat transfer from which amounted to 28.2 per cent of the total heat tion of the theory to any as advanced in the theory of combustion, as compared with 6.9 per cent from a great extent. of radiation from luminous similar non-luminous acetylene flame. flames. These data demonstrate the order of magnitude of Experimental Method and Theory

radiation from luminous flames and indicate that it should hold a position of greater importance among the methods of heat transfer than has been hitherto recognized by the majority of engineers.

According to the generally accepted theory of today, light and heat radiations are essentially the same kind of phenomena-namely, wave motions-differing only in length of wave. In the case of non-luminous flames it is believed these waves are set up a t the instant of chemical combination. The energy of the molecule can be divided into two classesthe kinetic energy, which is evidenced by that property of gases known as pressure, and an inner-molecular energy due to the movement of the electrons. It is believed that the chemical combination is brought about by an electronic rearrangement. Thus when hydrogen unites with oxygen to form water, this electronic rearranging sets up intensive vibrations of high frequency which result in the emission of energy in the form of radiation. This theory is borne out by the fact that inerts such as nitrogen in the flame show very slight emission of radiation. Additional support is derived from the fact that practically all of the radiation from an explosion occurs at the instant just previous to the instant of maximum p r e ~ s u r e . ~ In the case of the luminous flame there is, in addition to the radiation due to electronic rearrangement, common to both luminous and non-luminous flames, also the radiation from the incandescent carbon particles, which have been deposited in the flame by thermal decomposition. That there are solid particles in the flame is definitely shown by the following statement : l o Soret showed t h a t if a n image of the sun be focussed upon the glowing part of a hydrocarbon flame, the scattered light is polarized, and i t is therefore indisputable t h a t the luminous region is pervaded by a cloud of finely divided solid matter although the flame may not give rise to soot unless chilled.

Apparatus

With the exception of a few changes, the m e t h o d and apparatus e m p l o y e d in the experimental work are the same as used in the measurement of radiation from non-luminous flames.2 A galvanometer of greater sensitivity was used as well as water-cooled burner tips and better thermal insulation of t,ne thermopile chamber. Briefly outlined, the method was to burn an air and gas mixture of known composition under constant conditions and measure the radiation by some type of radiometer. To carry out this simple operation it was necessary to have constant supplies of gas and air, flow meters for the accurate measurement of gas and air, a combustion chamber not affected by temperature changes in the surroundings, a mixer which would give complete mixing of gas and air, 8 burner that would give the desired type of flame and not heat a t the tip, and a thermopile and galvanometer of sufficient sensitivity to measure the radiation from a small flame at a distance of 70 cm. The air and gas supply was measured with carefully Calibrated orifices, using inclined gages for accuracy. The burner tips were water-cooled, with a removable water jacket. It was found that the heating of a non-water-cooled burner affected the results appreciably. The combustion chamber consisted of a galvanized iron cylinder about 10 cm. in diameter by 30 cm. high, which was surrounded by a water jacket. About half way up the side of the combustion chamber and perpendicular to it was connected a section of 10-em. iron pipe 90 cm. long. This served as a chamber for the thermopile and was lagged with commercial magnesia insulation to protect the thermopile from temperature changes in the surroundings. The thermopile used was a twelve-junction copperconstantan pile, made by W. TV. Coblentz, of the Bureau

INDUSTRIAL AND ENGINEERING CHEMISTRY

6

of Standards. Used in conjunction with this was a Leeds b Northrup high-sensitivity galvanometer, mounted on a Julius suspension and observed through a telescope. The thermopile was calibrated against a Bureau of Standards radiation standard. The procedure followed was to note the galvanometerreading or adjust it to zero, light the flame, and observe it at regular short time intervals, keeping the gas and air rate constant, until the galvanometer deflection reached a maximum a t which time the flame was extinguished. Assuming the flame to be a point source, the radiation can be calculated from the calibration value and the deflection observed. Experimental Results

In view of the importance of radiation from luminous flames in industrial work and because of the meager amount of quantitative information a t hand, it was considered desirable to undertake the measurement of such radiation. Both luminous and non-luminous flames from four hydrocarbon gases were investigated, the most representative results being given in Table 11. Table I1 ~~

I

GAS USED

PRZiY' % of total Per cent

City gas Methane Ethylene Acetvlene

HEATRADIATED PROM THE FLAMES

... 50

NON-LUMINOUS

heat of K g . cal. per combustion g. mol.

...

7.0

l4:65

LUMINOUS

% of total

heat of Kg. cal. per combustion g. mol. 10.21 16.74

88.09

These data are summarized in the accompanying curves showing the relation between the per cent radiation and the per cent of primary air, expressed as per cent of the theoretical air required for complete combustion, for the three gases, methane, ethylene, and acetylene. The experimental work done on city gas flames was chiefly for developing experimental technic, and since there was some variation in the composition of the gas the results obtained with this gas are

not considered of special interest. As the experimental method was originally planned, the combustion chamber was to be filled with the inert gas nitrogen and all the air necessary for combustion was to be measured and supplied through the mixer. However, it was found to be practically impossible to produce a luminous flame with this arrangement, owing to diffusion of nitrogen into the flame, which would then go out. This idea was abandoned and the combustion chamber was filled with air under atmospheric pressure. Thus, in the plots of per cent radiation versm per cent primary air the per cent primary air means the amount of air supplied to the mixer. In addition to the primary air supplied, there was an

Vol. 19, No. 1

unmeasured amount of air which diffused into the flame from the combustion chamber. The luminous flames were obtained by burning the gas with zero primary air. The method of experimentation necessitated the use of small flames. This is to be criticized, chiefly because of the difficulty in obtaining a flame which was entirely luminous. The methane flame had a glowing cone with a blue base and the ethylene flame was similar, with a slightly larger glowing area. Undoubtedly, the values of radiation from luminous flames obtained with city gas, methane, and ethylene are somewhat low, and would have been higher if a more luminous flame could have been obtained. By a more luminous flame is meant here not only a flame in which the luminous region extended throughout the flame, but also a flame of greater depth. It has been pointed out that the greatest radiation should be expected from a flame of such a depth that the projected area of the carbon particles would be continuous. The results obtained with the acetylene gas flame are considered the most representative. It was only with this gas that a small flame could be obtained which was either smoky or which was entirely luminous. Acetylene is the least saturated gas with respect to hydrogen, and consequently the one which would be expected to burn with the greatest amount of cracking. (Even with acetylene higher values might have been obtained by using a deeper flame.) The importance of deposited carbon to radiation from a luminous flame is clearly shown by the effect of incraasing the primary air. The first result of adding primary air to a luminous flame is to reduce the cracking by furnishing sufficient air for hydroxylation or complete combustion of the gas. This, in turn, results in a smaller amount of carbon being deposited in the flame, which is evidenced by a smaller amount of radiation from the flame and a dcerease in luminosity. The curve of per cent radiation versus per cent primary air for acetylene demonstrates this point. Conclusions

Radiation from a truly luminous flame is a t least 25 to 30 per cent of the total heat of combustion of the gas. In going from non-luminous flame to a truly luminous flame the rate of heat transfer by radiation is increased a t least fourfold. With larger flames this relative increase may even be greater. From a consideration of the results in view of the saturation of the various gases used, it may be concluded that for luminous flames of gases burning with the same amounts of air the greatest amount of radiation will be given off by the flame of the least saturated gas. Radiation from luminous flames is a very important means of heat transfer, probably a more important means than many engineers have realized. The workman unconsciously realizes its importance when he insists that a bright or luminous flame will heat his charge more quickly than the blue or non-luminous flame. This work supports the suggestion of Lents concerning the inoculation of non-luminous flames with an illuminating agent-i. e., that the method of inoculation is an excellent one for increasing the rate of heat transfer from non-luminous flames-and will probably find many industrial applications. For example, industrial installations using blast-furnace or natural gas might apply this idea to an advantage. Bibliography 1-Von Helmholtz, "Die Licht und Wiirmestrahlung verbrannter Gase," Berlin, 1890. 2-Haslam, Lovell, and Hunneman, I n d . Eng. Chcm., 17, 272 (1925). 3--Pashen, Wiad. A n n . , 60, 409 (1893); 61, 1 (1893); 6% 209 (1894). 4--Callender, J . G a s Lighting, 111, 644 (1910). 5-Hunnema11, Chemical Engineering, Thesis, M. I. T.,1923. 6-Lovell, I b i d . , 1924. 7-2. Ver. deut. Ing., 68, 1017 (1924). 8-Wdrme, 49, 145 (1926). 9-David, Proc. Roy. Soc. (London), A86, 575 (1911). 10-FueZ Science Practice, 4, 11 (1925).