heterogeneous flash initiation of thermal reactions - American

---

March, 1959

HETEROGENEOUS FLASH INITIATION OF THERMAL REACTIONS

433

HETEROGENEOUS FLASH INITIATION OF THERMAL REACTIONS' BY L. S. NELSON AND J. L. LUNDBERG Bell Telephone Laboratories, Incorporated, Murray Hill,New Jersey Received November 17, 1068

Heterogeneous flash initiation is a widely applicable technique for initiating thermal reactions. The only requirements are: (1) a very intense light source, in our case a flash lam which produces light pulses of several milliseconds duration, and (2) a finely divided light-absorbing material (particles, figments or foils) suspended in (3) a reasonably transparent poor thermal conductor. The radiant flux (near ultraviolet, visible and near infrared) is mainly converted to heat a t the absorbing surface, causing local temperature rises which depend on the incident light intensity and the rate of energy dissipation. For materials of smallest dimension near 10 p these temperature rises approach 2000" in condensed phase insulators and 5000" ~ T vacuo L or low pressure gases for the experimental conditions used. Calculated temperature rises agree with experiment. Thermal reactions have been initiated at solid surfaces surrounded by a vacuum, gases, liquids and transparent polymers, resulting in vaporization, melting, yrolysis, explosion and other rapid reactions. Because of the short initiation times required, flash heating should find appication analogous to flash photolysis and shock tube techniques for producing labile intermediate species in rapid reaction studies.

Introduction Earlier woik by the authors2 has shown that tiny dark particles suspended in a transparent polymeric matrix can be heated to very high temperaiures by black-body absorption of intense flashes of light. During this heating, the temperature of the matrix rises sharply, and pyrolysis occurs. Here we shall show that this flash heating is a rather general phenomenon which may be used to initiate many thermal reactions at temperatures and on a time scale not easily attainable by other means. The requirements for such initiations are simply (1) a very intense light source, in our case a capacitordischarge flash lamp, (2) a finely divided dark material in the form of particles, filaments or foils suspended in (3) a fairly transparent poor thermal conductor. The thermal reactions may occur in the insulating matrix, as in the case of the polymer pyrolysis, or may involve the finely divided material, or some combination of the two. We have used the technique to obtain intense rapid heating a t solid surfaces surrounded by vacuum, gases, liquids and transparent polymers. I n some cases we have explosively evaporated tungsten powders and wires in vacuo with subsequent deposition of heavy films, an indication that instantaneous tempera tures approaching several thousand degrees centigrade are readily obtainable in our flash times of several milliseconds.2b It is thuq possible to initiate flash pyrolyses, carbon format ion, explosions, evaporations, meltings and other thermal processes in systems without discrete photochemical absorption in a manner somewhat parallel to the flash photolysis technique3 which has been applied successfully to the preparation of labile molecules for kinetic spectroscopy4 or rapid mass spe~trometry.~In gases, the method produces effects similar to shock tube pyrolyses6 but (1) Presented before the American Chemical Society, Chicago, September, 1958. ( 2 ) (a) J. L. Lundberg and L. S. Nelson, Nature, 119, 367 (1957); (b) J. L. Lundberg, L. S. Nelson and M. Y. Hellman, "Proc. Third Conference on Carbon," Buffalo, N. Y., 1957,in press. (3) G. Porter, Proc. R o y . Soc. ( L o n d o n ) , A200, 284 (1950); G. Haraberg and D. A. Ramsay, J. Chem. Phys., 20, 347 (1952); R. Marshall and N. Davidaon, ibid., 21, 659 (1953). (4) R. G. W. Norrish and R. A. Thrush, Quart. Reu. ( L o n d o n ) , 10,

149 (1956). ( 3 ) G . B. ICistialiowsky and P. €1. Rydd, J. A m . Chem. Soc., 79, 4825 (1957). (6) E.F. Creene, R.L. Taylor and W. L.Patterson, Jr., THISJOTJRNAL, 68,238 (1968).

with a considerably larger instantaneous volume of labile intermediates, which should simplify kinetic measurements and end-product analyses. Procedure .-The system for irradiating the samples is shown schematically in Fig. 1. We used spiral flash lamps' very similar to Christie and Porter's Type III.8 The lamps were made of fused silica and filled with xenon at about 70 mm. pressure. In order to obtain maximum incident flux, the lamp was surrounded by an aluminum cylinder coated inside with ma nesium oxide. This type of flash lamp emits very nearly a #at continuum extending from the near ultraviolet through the visible into the near infrared region of the spectrum.9 The lamp output between 2000 and 4000 A. was shown by uranyl oxalate actinornetrys*"J to be 1019-1020 quanta flash-' cm.+ for flashes of several milliseconds duration,2bdepending on the capacitance used. The lamp was powered by a commercial photographic capacitor bank and power supply11 capable of charging to 4 kv. capacitances of 36 to 1296 pF., variable in 36 p F . steps. The materials to be irradiated were enclosed in No. 7740 chemical Pyrex tubes of 1/2 in. diameter, which were flashed while suspended along the spiral axis of the lamp.

Experimental Results In this section we shall describe a number of experiments which indicate the applicability of heterogeneous flash initiation to thermal processes. We shall discuss them according to the phase of the insulating matrix. All flash energies E have been calculated from E = 1/2CVz,where C is the capacitance and V is the charging voltage (always 4 kv.). Solid-Vacuum Experiments.-The simplest thermal reaction was the evaporation of a finely divided dark-colored material suspended in vacuo. With one flash at 5184 j . we evaporated sieved Ag, C, Cr, Mo, Ni, W and Zn of particle diameter below 100 p suspended on the walls of an evacuated Pyrex tube. The lower boiling metals, such as Zn and Ag completely gasify a t the lower energies of 2592 j ./flash, laying down heavy, smooth mirrors, while the higher boiling metals such as Mo and W, are more difficult to evaporate in powder form and require about 10,000 j./ffash for heavy depositions of metal. Tungsten wire of 7.5 diameter, when tangled into a wool and suspended in an evacuated '/2 in. diameter Pyrex tube, evaporated somewhat a t 2592 j./flash and explosively a t 5184 j./flash. Amorphous and graphitic carbon powders (< 100p diameter) also evaporated quite easily on flashing at 5184 j./flash with deposition of heavy mirrors of carbon, and considerable internal crazing inside the Pyrex tube in which it was suspended. (The crazing is quite common in these experiments, occurring a t the higher energies per flash with many materials; it never has caused shattering or leakage of the tubes, however.) (7) Prepared by the General Electrio Co., Lamp Division, Cleveland, Ohio. (8) M. I. Christie and G . Porter, Proc. R o y . Soc. (London). A212, 390 (1952). (9) M. I. Christie and G. Porter, ibid., A212, 398 (1952). (10) G . S.Forbesand L. J. Heidt, J. A m . Chem. Soc., 6 6 , 2363 (1934). (11) Obtained from the Edgerton, Gerrneshausen and Grier Corp., Boston, Maas.

434

L. S. NELSONAND J. L. LUNDBERG

FVSED S,L,CA

TEST

T ~ B E

S A ~ P ~ E

,026:LA

Fig. 1.-Circuit

and lamp for flash heating.

WAVELENGTH IN

00 ANGSTROMS.

Fig. 2.-Absorption spectra of mineral oil samples after one flash with various capacitances charged to 4 kv. Foils of Ni and Perminvar (5-10

thick) suspended i n j./flash) and evaporate rapidly a t higher energies (-5000 j ./flash). At intermediate energies etching occurs, and microscopic structure becomes apparent on the surface of the metal. When the finely divided material is of organic nature, various thermal reactions can be initiated in the flash. An exploratory joint experiment with the Bureau of Mines on the flash heating of powdered coals sealed off in Pyrex tubes has shown some interesting pyrolytic results.12 The hydrocarbons formed mere of four carbon atoms or less, and in very different distribution than observed in conventional I n addition, C2H2, methylacetylene, CSZ and &yrO1yses CN were observed. Obviously this is a pyrolytic technique requiring more investigation. Solid-Gas Experiments.-By the technique of flash heating, we have decomposed gaseous CH, (Phillips Research Grade) by suspending 25 m. of 7.5 p diameter black W wire as a loose wool in a in. diameter Pyrex tube filled with 285 mm. of the gas. With one flash a t 2592 j. carbon was deposited, and, as shown by gas chromatography on a silica gel column, H2, C2HB,C2Hl and C2H2were formed. A control sample of CH, flashed without the wire showed no change. By comparison of the chromatographic peak areas p

vacuo melt when flashed at lower energies (-2000

.

(12) Personal communication from Dr. M. D. Schleainger.

Vol. 63

with known mixtures, i t was found that 3 micromoles of HZwas formed per meter of wire. Assuming that for every mole of HZproduced, one-half mole of CH, was decomposed, a t the pressure used a cylinder of CHI a t least 0.36 mm. in diameter was pyrolyzed. This means that the wire influences about 400 times its own volume in the gas phase. Since methane does not absorb radiation in the visible or near ultraviolet, this is an example of heterogeneous “sensitization” which has been used to convert a short pulse of predominantly visible radiation into chemical energy, somewhat like the action of homogeneous photochemical sensitizers. We also have initiated explosions in the non-absorbin systems CHI-02 and Hr02. Metallic initiators such as \$ and Mo could not be used, because of a rapid oxidation of the metal in preference to the explosion. However glass wool evidently absorbs sufficiently over the wave length range of the flash to rise to the temperature for initiation of the explosion, although only a t the highest flash energy (10,386 j./flash). In 1/2 in. diameter Pyrex tubes, loose 5 mg. skeins of ordinary laboratory grade Pyrex wool ( 4p fiber diameter) were closed off with 100 mm. of stoichiometric mixtures of CHI or H2and 0 2 . One flash triggered the explosion, the occurrence of which was detected by condensing the products in liquid nitrogen. After the flash, some of the more densely packed glass fibers were fused into coarser fibrous networks. Solid-Liquid Experiments.-We have flash pyrolyzed a commercial mineral oil containing suspended sieved Mo of diameter below 74 p, C of diameter below 36 p or W wool of 7.5 p diameter. The oil was placed in a 1/2 in. diameter Pyrex tube with 0.03% by weight of the finely divided material. Above a threshold flash near 7500 j . the mineral oil turned an inky black (as compared to a grey cast before flashing in the case of the two powders), effervesced a small amounot, and exhibited a bluish fluorescence when excited by 3661 A. mercury radiation. Electron micrographs of the new carbon particles showed chain-like aggregates of spherical particles approximately 100 d. in diameter. Below the threshold energy, no change in the suspension could be detected visually. Outgassing to a residual pressure of 10-6 mm. had no effect on these experiments. A series of mineral oil samples pyrolyzed with carbon particles by flashing a t different energies was filtered overnight in air through a fine filter paper. Absorption spectra of each clear filtrate were recorded spectrophotometrically, in 10 mm. thickness as shown in Fig. 2. The lower curves marked 756 and 864 pF. are very nearly coincident with a control mineral oil sample which had been mixed with the sieved carbon, allowed to stand in contact with the carbon as long as the other samples, but filtered without flashing. The absorbance curves shown in Fig. 2 are quite similar to curves obtained from mineral oils which have undergone high-field electrical degradation .13 The absorbance increase in this case is thought to result from formation of large unsaturated molecules. In com arison, no degradation of the type seen above could be g u n d in purified liquid n-octadecane suspensions of the same materials, even with repeated flashes of tmhe highest energies. This probably is due to cooling by microscopic vaporization even though no bubbles were visible to the eye several seconds after the flash. Phase changes in the matrix which occur below its decomposition temperature will very likely prevent pyrolysis by this technique. The case of flashing 0.03% by weight of zinc powder (