AF.OSR TR 59-185 THE UN I VE RS IT Y OF MI CHI GAN COLLEGE OF ENGINEERING Department of Chemical and Metallurgical Engineering Technical Report COMBUSTION EXPERIMENT WITH A 50,000-CURIE GOLD SOURCE Stuart W. Churchill Alexander Weir, Jr. Leroy F. Ornella Roy L. Gealer UMJRI Project 2288 under contra'cti:w''twth':: DEPARTMEN1T OF THEBEA-IR FORCE AIR RESEA.RCH A.ND DEVELOPM0ENT COOMMANT coNTRACT No'0 AF:i.-18(6oo00) -1218 WASHINGTON, D.C. administered by: TtIE UNIVERSITY OF MICHIGAN RESEARCH INSTITUTE ANN ARBOR November 1959

*!l~ ~ v,~

iii PREFACE This research was supported by the United States Air Force through the Air Force Office of Scientific Research, Air Research and Development Command. Personnel of the University of Michigan Aircraft Propulsion laboratory assisted in the construction of the equipment and in the experimental work. Mr. C. C. Palmiter and other personnel of the University of Michigan Radiation Control Service assisted in the design of the experiment, directed the transfer of the radioactive gold from the Materials Testing Reactor to the laboratory, and monitored the source during the experiment. The radioactive gold was transported from the Materials Testing Reactor to Ann Arbor by an Air Force flight arranged for by the Offices of Scientific Research.

iv ABSTRACT The data obtained in a previous experiment in which the flame zone and preflame.zone of premixed propane and.air were both irradiated with-.a 12000-curie gold source have been reworked and reinterpreted for journal publication FInd. Eng. Chem. 49, 1419-22, 1423-28 (1957)3. The previously reported effects of irradiation on the flame speed and emission spectra of the flames were reconfirmed but the uncertainty in the recomputed OH-rotational temperature was found to be greater than the observed variation with source strength and hence was not determinable from these data. A new experiment was designed to irradiate the flame and preflame zones separately under essentially the same conditions. An accidental explosion in the test chamber shortly -after installation of the radioactive gold damaged the burnerand the optical transmission system used for remote observation of the flame, preventing.collection of the intended data. The details of the.experimental design and the history of the experiment are nevertheless reported.as a guide to future.research of a similar nature.

TAB2E OF CONTENTS Page INTRODUCTION 1 EXPERIMENTAL APPARATUS 3 A. Source and Source Containers 3 B. Burner 5 C. Test Tank 5 D. Shielding for Personnel 6 E. Optical System 6 F. Instrumentation 6 HISTORY OF TE EXPERIMENT 7 CONCLUSIONS 9 BIBLIOGRAPHY 11 LIST OF FIGIURES 1. Previous Combustion Experiment 2 2. Experimental Arrangement 5 3. Disintegration Schemes of Au-198 and Au-199 6 4. Installation of Gold-Coated Slabs in Carriage 12

vi BIBLIOGRAPHICAL COJTRL SHT 1. Originating agency and monitoring agency: 0. A.: University of Michigan, Ann Arbor, Michigan M. A.: Office for Advanced Studies, Air Force Office of Scientific Re-search 2. Originating agency and monitoring agency report number: 0. A.: JUMRI Technical Report 2288-9-T M. A.: AFOSR Technical Report 59-185 3. Title- and classification of title: COMBUSTION STUDIES WITH A 12000-CURIE GOLD SOURCE (UNCIASSIFIED) 4. Personal authors: Stuart W. Churchill, Alexander Weir, Jr., Leroy F. Ornella, Roy L. Gealer 5. Date of report November, 1959 6. Pages: 20 7. Illustrative material: 4 figures 8. Prepared for Contract No.: AF 18(600)-1218 9. Prepared-: -for Project No.: Task 37507 10. Security classifi-caton- UNCIASSIFIED 11. Distribution limitations: None 12. Abstract: The data obtained in a previous.experiment in which the flame Zone' and preflame zone of premixed propane and air were both irradiated with a 12000-curie;gold source have been reworked and reinterpreted for journal publication lind. Eng. Chem. 49 1419-22, 1423-28 (1957)]. The. previously reported effects of irrdiat-ion on the flame speed and- emission spectra of the flames were reconfirmed, but the uncertainty in the recomputed OH-rotational tempe:rature was found to be greater than the observed variation with source strength and hence was not determinable from these data. A new experiment was designed to irradiate the flame and preflame zones separately under essentially the same conditions. An accidental explosion in the test chamber shortly after installation of the radioactive gold damaged the burner and the optical transmission system usd for remote observation of the fle, preventing collection of the intended data Thedetails of the experimental design and the history of the experiment are nevertheless reported as a guide to future.research of a similar nature.

INTRODUCTION In the course of investigations of the utilization of gross fission products at the University of Michigan, flames were irradiated with gold and palladium sourceso These investigations, which were carried out at low intensities, indicated some effect upon the flame stability (1) and no effect upon flame speed (2, 3). Consequently an experiment was designed to study the effect of more intense radiation upon flames. The results of this study were reported in a Technical Note (4) and subsequently in two journal articles (5,6). The data were reworked and reinterpreted in the preparation of references 5 and 6, and these two articles should be considered to- supercede completely the Technical Note. The source in the.above experiment consisted of 30 grams of irradiated gold with an initial strength of approximately 12,000 curieso The gold was in the form of coils of 0.005-inch wire. Premixed propane and air were passed through a compacted mass of these coils and burned above.a screen as shown in Fig. 1. Thus both the preflame mixture and the flame zone were irradiated. The ionization produced by beta radiation exceeded that from gamma radiation and was estimated to be in the order of 3 x 1015 and 8 x 109 ion pairs/(cm)5 (sec)(curie of Au-198 or Au-199) in the region of the gold and in the flame zone, respectively. The spectral emission at a series of elevations through a flat flame and the rate of propagation of bunsen-type flames werel measured for a series of propane-to-air ratios as the source strength decreased. Increases in the rate of propagation of up to 50%, and increases in the-emission due to CI and C2 of up to 30% and 150%, respectively, and vertical displacement of the

2 UPPER BASKET SCREEN FLAME BASKET.-SCREEN SCREEN CLAMP RING IRADIOACTIVE GOLD CLAMP STUD REGION C OOLING WATER GROOVE \-SCREENED SNAP-RING CHANNEL TO GROOVE BURNER TUBE COOLING WATER OUT COOLING WATER IN TO PACKING GLAND Fig. 1. Previous combustion experiment.

-3maximum in the spectral emission were observed to result from the irradiation. OH-rotational temperatures were computed from the spectral emission and are reported in the Technical Note (4). However, subsequent reanalysis of these computations revealed several significant errors. The uncertainty in the recomputed OH-rotational temperatures was greater than the variation with source strength. Hence these values are not considered to have significance and were not included in the journal articles (5,6), which supercede the Technical Note. It was not possible to determine- from the- above experimentwhether the observed effects were primarily due to irradiation of the flame zone, or of the preflame mixture, or of both. This report presents the results of a further experiment which was designed to permit separate irradiation of different zones of the propane-air and flame mixtures. It was intended to attain essentially the same maximum level of ionization -as in the previous experiment and again to measure the emission spectra of the flame, the flame speed, and such performance characteristics as blow-off velocity and specific air impulse. Unfortunately an explosion occured inside the experimental apparatus soon after installation of the irradiated gold. Damage to the optical system prevented attainment of spectral data and flame speed, and damage to the combustion apparatus itself produced a skewed flame.and anomalous performance data. The presence of the radioactive source made adjustments impossible. Therefore no experimental data are presented in this report. The design and history of the experiment are nevertheless reported herein for the- guidance of future experimenters in the- field.

E2E]4E2IAL APPARATUS The experimental equipment consisted primarily of a premixed propane-air supply and metering system,.a burner head a movable-.source of radiation, a vacuum tank and system to control pressure at sub-atmospheric levels.an optical system to transmit an' image of the flame outside the source region, instrumentation to analyze the image of the.flame, and radiation shielding0 The overall experimental arrangement is shown in Fig. 2 and the pertinent details are discussed below. A. Source and Source Containers Gold was selected as a source material in the- previous experiment because the half-life and strength obtained by irradiation are more. suitable for combustion studies than.those from other sources of beta.radiation. As indicated in the disintegration scheme in Fig. 3, the half-life of Au-198 is 2.69 days while that, of Au-199 is 3.15 days (7). This average half-life of about 3 days allows a 24-hour experimental run to be completed with only a 20% decrease. in radioactivity. The gamma radiation which is emitted from both Au-198 and Au-199 is of course somewhat disadvantageous because of shielding problems and the possibleeffect upon the combustion process, but is only a minor contributor to the formation of ion pairs (5). Gold was selected for this experiment because of these reasons and because of the desirability of comparison of the results with the previous experiment. A detailed analysis of cost, geometry, and mechanical operation indicated that the objective of separate irradiation of a preflame mixture

EXPERIMENTAL ARRANGEMENT HEAD LEVEL FOR. RADIOACTIVE GOLD SOURCE w w 4 o z C o 0 w H 0I I GROUND LEVEL CONCRETE FLOOR TEST TANK P. SOURCE RING —-, BURNER/I Fig 2.Experimental arrangement.

Au-198, HALF LIFE 2.69 DAYS \ - (0.96 rne.v. ) - (0,41 m.e.v.) Au -199, HALF LIFE =3.15 DAYS -,8(0.32 me.v. ).r(.158 me.), (67m%, ) (0.208me*), 33% r(0.05 onmauv.) Fig. 3. Disintegration schemes of Au-198 and Au-199.

and a l-inch-diameter flat flame, at levels corresponding to an energy absorption of 50 x 108 mev/(cm)3(sec) could be best attained with a cylindrical ring; 5 inches high and 8- inches in diameter composed of 52 1/16-inch-thick by i-inch wide by 2-7/8-inch high aluminum slabs coated on one side with 0.00266 inch of gold. The gold coating totalled 60 grams and the initial source strength was expected to be about 75000 curies. The slabs were- held in an aluminum carriage consisting of an outer shell, a grooved base, and an inside retaining wire. This car~iage, ring was then fastened to a hydraulic jack together with a dummy ring spaced 5 inches above the source ring as indicated in Fig. 2. In the uppermost position the source-ring irradiated the flame zone, and the burner head shielded the preflame mixture. In the downmost position the source ring irradiated the preflame mixture, the burner head shielded the flame zone, and the dummy ring preserved the same geometrical surroundings for the flame. The dimensions of the gold-plated slabs were determined by the permissible dimensions of the capsules to be installed in the Materials Testing Reactor as well as by the desired experimental conditions. For irradiation the slabs were placed in four aluminum capsules 3 inches long with an inside diameter of 7/8 inch and an outside diameter of 1-1/8 inch. After irradiation the capsules were placed in a lead shipping container 15 inches high and 141 inches in diameter with a central cavity just big enough for the four capsules. The lead container was clad in 1/8-inch -steel.

-8B. Burner The burner head consisted of two brass blocks 4 inches in diameter and 1- inches thickj:with a central passage 1 inch in diameter connected by a 4 —inchlong by 1-inch-inside-diameter brass tube.with a wall thickness of only 0.001 inch. The upper block contained a sintered bronze plug 1 inch in diameter at the outlet, and water channels for cooling of the plugo This design permitted a high percentage of the radiation to pass through the pref lame mixture but shielded the flame zone when the source ring was in the lower position and also shielded the preflame mixture effectively with the source.ring in the upper position. The cooling water lines and the leads from a thermocouple attached to the sintered plug passed out through a length of 3/4-inch pipe connected rigidly to the burner head and serving as a handle for remote installation and removal of-the burner head. The propane-air mixture was ignited by a spark between two 1/8-inch rods connected to a high-voltage transformer. The rods were mounted on a swivel connected to a hydraulic jack which could be used to swing the ignitor away from the burner after ignition. C.. Test Tank The burner head and source were located in -a tank 30 inches in diameter and 9 feet high. A vacuum pump connected to the exhaust line from the tank was capable of maintaining pressures as low as 3 in. HIg. abs. during a combustion experiment. The removable cover of the tank contained an aluminum rupture diaphragm 15 inches in diameter to prevent destruction of the tank in

the event of an accidental explosio- of accumulited gases. The tank could be filled with water to protect personnel from radiation during installation of the: gold source. Storage tanks were connected to the- experimental tanks to retain this water. D. Shielding for Personnel The- test tank -was located in a. concrete pit as shown in Fig. 2. The walls of the pit, the earth, and.a 12-inch-thick concrete wall isolating the experimental apparatus from the laboratory area provided a.dequate shielding for personnel after lead sheets were placed at.a few specific locations indicated by radiation measurements. Personnel entered the test area after installation of the source only when the test tank was filled with water. 3. Optical System The image of the flame left the tank horizontally through a i-inch thick quartz window, was reflected downward by a flat mirror, then upward by a 4I-foot-focal-length paabolic mirror to a flat mirror, then to a 12-foot focal-length parabolic- mirror'which focused the image on the.entrance.slit of the monochrometer. The.entire optical system was enclosed in tubing to prevent the entrance of stray light. F. Instrumentation The flow rates of propane and air were determined with capillary meters. A thermocoupe in the sntered plug in the burne head provided.a

-10measure of the temperature of the preflamne mixture. A manometer indicated the pressure in the test tank. A Jarrel-Ash monochrometer was positioned outside the isolation cell with its entrance slit at the focus of the image of the flame supplied by the optical system. The monochrometer scans the spectrum -at a speed set by the operator and the monochromatic light falls on a photomultiplier tube-at the exit slit. The photomultiplier was connected to a Brown recorder through an amplifier and the result was a plot of relative intensity versus time-, with time proportional to wave number. The monochrometer was movable in a direction perpendicular to the plane of the image of the flat' flame so that the flame could be scanned vertically, producing a recording of the variation of the intensity at a particular wavelength with height. The monochrometer could be swung out of the light path to permit the positioning of'a photographic film holder at the focus of the flame image. Such photographs were intended to provide a means of determining the flame speed. HISTORY OF THIE EXPERIMENT After extensive rehearsals of the experiment in the absence of a source, the gold-plated aluminum slabs were sent to the Materials Testing Reactor, Idaho Fall, Idaho for irradiation. After irradiation for approximately five days the capsules were removed from the reactor, installed in the lead container and flown to Ann Arbor. Upon arrival at the- laboratory the slugs

were removed from the lead con-tainer and capsules, and installed in the source ring under water, using specially constructed tools for remote handling from the top of the tank. This operation is show in Fig. 4. Of the 52 slugs 47 were successfully installed in the source ring. The other 5 were accidentally dropped to the bottom of the tank in inaccessible positions where they remained throughout the experiment. The water was then drained from the tank, and the tank and burner dried with air. Propane was introduced and ignited. Shortly thereafter a mild.explosion occurred in the test tank blowing out the rupture disk. The cause of this accident is believed to be.as follows. The intense irradiation from the radioactive gold increased the flame speed to the- point that combustion occurred in or immediately adjacent to the sintered bronze plug. Solder holding the sintered plug in place melted due to the resultlng excessive temperatures in the burner head, resulting in leakage of gas around the-edge Qof the bronze plug. Since the propane-air mixture flowed preferentially through this crack, theregion near the center where the spark gap was located probably was too lean to ignite. Unburned gas therefore accumulated in the test tank until eventual igfiition created excessive pressure in the tank. The -explosion resulted in misalignment of the mirrors in the optical piping to the laboratory area. It proved impossible to realign the optical system remotely, and it was impossible to enter the test pit even if the gold plated slabs were removed from the source carriage because of the inacessible slabs which were dropped to the bottom of the test tank.

lJ; 4:H Ii 0 0 o H 0 cP lCD IJ. (3 CD HM -i-)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ —---

-13The top section -of the burner was removed from the test tanik for repairs. However, the- rapid decrease in source strength with time necessitated hasty work, and radioactivity of the burner head due to contamination by the shielding water necessitated remote handling. These makeshift repairs proved to be unsuccessful in that the data which were- subsequently obtained for blowoff limits, e tc., were erratic and uninterpretable. It is presumed that a leak persisted in the thin tube connecting the upper and low parts of the burner. CONCLUSION The only conclusion that can be stated relative to the.effect of radiation of the flame- is that the radioactive gold did produce gross effeqots such as were observed in the earlier experiment. Other than the factors and events which lead to the.accidental explosion, the design of the experiment appears to have been satisfactory, The strength of the source turned out to be of the order of 50,000 curies, as predicted, despite early removal from the Material Test Reactor duo to reactor operating difficulties. The shielding provided for personnel during the handling of the gold.and during the: experimentation also proved to be adequate. A different tool for remote. handling of the slabs to minimize the chance of dropping a slab is certainly desirable. The use of high-temperature materials in the burner head, and extreme care during the.establishment of a flame are obviously nevcessary due to the abnormal behavior of flames at extreme levels of irradiation. Recent examination of the- gold-plated.slabs used in the.exeriment reveals a significant but not irntolerable'amount of residual radioactivity due

,614to impurities in the aluminum. It is concluded that these slabs, which wereexpensive to fabricate, could be irradiated and used again in the same or a similar- experiment,

-BIBLIOIRAPY 1. Cullen, R. E., and M. E. Gluckstein, "Effect of Atomic Radiation on the Combustion of Hydrocarbon Air MixturesT", Fifth Symposium (International) on Combustion, p. 569, Reinhold Publishing Corp., N' Y. (1955). 2. Morrison, R. B., R. E. Cullen, and Alexander Weir, Jr., "Utilization of Gross Fission Products; Performance of Combustion Engines under the Influence of Radiation - Jet Engines", The University of Michigan Engineering Rsearch Institute, A.E.C. Contract No. AT(11l-1)162, Progress Report No. 2, (January 31, 1952). 3. Morrison, R. B., R. E. Cullen, and Alexander Weir, Jr., "Utilization of Gross Fission Products; Performanbe- of Combustion Engin'es under the Influence of Radiation - Jet Engines", The University of Michigan Research Institute, A.E.C. Contract No. AT(ll-1)162, Progress Report Noo 3, (June 30, 1952). 4. Churchill, S. W. Alexander Weir, Jr., L. F. Ornella, R. L. Gealer, R. J. Kelley and M. E. Gluckstein, "Combustion Studies with a 12000-Curie Gold Source", The University of Michigan Research Institute, Report 2288-6-T, AFOSR-TN-56-17, Ann Arbor, Michigan (December 1955). 5. Churchill, S. W., Alexander Weir, Jr., R. L. Gealer and R. J. Kelley, t"Rate of Propagation of Propane-Air Flames Irradiated with a 10,000-Curie Gold Source-", Ind. Eng. Chem. 49, 1419 (1957). 6. Weir, Alexander, Jr., S. W. Churchill, L. F. Ornella, M. E. Gluckstein, "'Emission Spectra of Propane-Air Flames Irradiated with a 1000-Curie Gold Source", Ind. Eng. Chem. 49, 1423 (1957). 7. "Nuclear Data", Nat. Bur. Stds. (U.S.) Circ. 499, Suppl. 3, Washington, D, C. (June, 1952).

UNIVERSITY OF MICHIGAN 3 901111111111151111111111111 III 36ilI 3 9015 02828 3763