54 MECHANICAL METHOD FOR ACTIVATION A MECHANICAL METHOD FOR THE ACTIVATION OF FAST REACTIONS BY T. H. BULL* AND P. B. MOON Department of Physics, University of Birmingham Received 2nd February, 1954 By relatively simple mechanical means, pulsed beams of molecules moving with speeds up to about a kilometre per second can be obtained. Heavy atoms and molecules moving at such speeds carry energies of the order of I0 kcal/mole and may react with other molecules with which they collide. Preliminary experiments have been made on the formation of CsCl by the impact of CC14 on Cs. During the past few years, a technique has been developed by which gas mole- cules at low pressure are swept up by the tip of a rotating blade and are flung off with the speed of the tip added to their thermal velocities.Since the speed of the * Present address: 1. C. I., Ltd., Butterworth Res. Lab., Welwyn, Herts.T. H . BULL AND P . B. MOON 55 tip may reach 105 cmlsec, the kinetic energy given to heavy molecules can be very substantial; for a molecular weight of 200, 10s cmlsec corresponds to 105 J/mole or about 24 kcal/mole. If such a fast molecule collides with some other molecule, it is possible that chemical reaction may ensue and, at first sight, it seems attractive to surround the rotor with a mixture of two gases and look for reaction products. It turns out, however, to be extremely difficult to distinguish any such effect in the presence of reactions occurring by the ordinary thermal mechanism in the gas or at the walls of the vessel ; it seems better to select a more-or-less collimated beam of one type of molecule and to direct this beam into another vessel where it meets a stream of the other molecules. Products of reactions genuinely due to high-speed impact must then be formed in the localized region of space where the beams cross ; more- over, they must be formed at identifiable instants of time, since the high-speed beam arrives in pulses, one for each revolution if the rotor carries a single blade.If both the projectile and the target molecules are heavy, but not so complex as to possess many degrees of freedom, a useful picture of the situation is obtained by ignoring thermal energies in comparison with the kinetic energy &qv,.2 corre- sponding to the mass ml of the projectile and the velocity vT of the rotor tip.Con- servation of momentum reduces the available energy by the factor m2/(ml + where m2 is the mass of the target molecule, the remainder going to the kinetic energy of the complex. Since this factor is typically of the order of Q, the avail- able energy is rather low for the activation of chemical reactions; in addition to choosing heavy molecules for study, it is clearly desirable to investigate examples where the activation energy is likely to be low. We owe to Prof. H. W. Melville the suggestion that reactions between alkali metal atoms and halogen compounds should be studied, and such reactions are attractive for another reason : the alkali metals and their compounds are sensi- tively and rapidly detectable by surface ionization on a hot filament.Though the technique requires further development before results of chemical significance can be obtained, it is felt that an account of the methods and preliminary results may be of interest. EXPERIMENTAL The apparatus, constructed of Pyrex glass, is shown in outline in fig. 1. Descriptions of the rotor and of the method of driving it have been given elsewhere,ls 2 and for present purposes it will be sufficient to say that the rotor blade was of aluminium alloy and that the effective area of the tip was about 0.25 cm2. At the highest speed used in these particular experiments (1500 revlsec or 6 x 104 cmlsec), the tip sweeps a volume of 15 l./sec. By a continuous-flow method, the pressure of CC14 in the rotor vessel was adjusted to be about 2 7 x 10-4 mm corresponding to a kinetic theory free path of about lOcm and to a vapour density of 2.5 x lO-gg/cm3. The mass of CC14 swept up by the tip each second was thus of the order of 4 x 10-5 g.A small fraction of these fast mole- cules passed through the slit s, through the region R (into which a stream of caesium atoms could be injected at a later stage of the work) and into an open-ended cylinder K which received electrons emitted by the hot tungsten filament F. As each pulse of mole- cules passed through the electron stream, some became ionized and caused an increase in the current to the anode. These pulses of current could be amplified and displayed on an oscilloscope with a time-base synchronized to the frequency of the rotor. Typical records are shown in fig.2a; the sinusoidal lower trace is a time-marker at twice the frequency of the rotor. Such records not only demonstrate the pulsed nature of the high-speed molecular beam ; they also give the actual time of flight of the pulse from the rotor to the detector. Table 1 shows the measured speed of the middle of the pulse compared with the speed of the rotor tip. The molecules in the pulse have a higher speed than the rotor tip, the roughly constant difference being about what would be expected from the thermal velocities with which they leave the tip. The next stage in the experiments was to introduce caesium vapour into the region R. This was done by breaking, with a magnetically-operated device, a tube containing caesium56 MECHANICAL METHOD FOR ACTIVATION metal which was surrounded by an oven 0.The caesium pressure was adjusted so that a CC14 molecule crossing the caesium stream would have an appreciable chance of making a collision but would be unlikely to make more than one collision with a caesium atom. - 0-- k CgE!.g-’ @pjRr Hor irontol Section FIG. 1. The caesium vapour was localized as far as possible by a liquid-air trap having as its lower end a hollow ring C through which the CC14 beam passed on its way to the detcctor. The cylinder K was now made negative with respect to the tungsten filament, which acted as a surface ionization detector both of caesium and of any CsCl that might be formed. TABLE 1 rotor CC4 “ positive ion ” tip speed pulse speed pulse speed ratio cmlsec cm/sec cm/sec 1.52 x 104 3-5 x 104 4.7 x 104 1.3 2.92 x 104 4.5 x 104 4.9 x 104 1.1 4-33 x 104 6.1 x 104 5.2 x 104 0.87 6-52 x 104 8.2 x 104 6.7 x 104 0.81 RESULTS Fig.26 shows the pulses that were observed. They are undoubtedly due to caesium ions leaving the tungsten filament, but whether they represent the arrival of Cs atoms or of CsCl molecules cannot be determined dircctly. From the times at which these positive ion pulses are observed, relative to those at which the CCI4 pulses reach the detector, it is possible to deduce the mean velocity with which the pulse of Cs or CsCl travels from R to F. Table 1 shows the results obtained, the last column giving the ratio of the pulse velocity to that of the carbon tetrachloride. If the caesium were in the form of atoms resulting from elastic collisions, this ratio should, from simple dynamics, be 1.07; for complerely irtelastic collisions it would be 0.54.Caesium chloride molecules would have velocities dependent on what fraction of the encrgy released in the reaction appears as kinetic encrgy. Detailed analysis of these results shows them to be consistent with the supposition that the observcd pulses are due to CsCl molecules formed in a reaction in which thc release of kinetic energy is about 3 kcal/molc ; but it is entirely possible that the pulses contain some Cs atoms, either projected without chemical change by the CC4 moleculesFIG. 2b. [To face page 56T. H . BULL AND P . B . MOON 57 or due to a small amount of caesium vapour entering the rotor chamber and being returned as a high-speed beam along with the main beam of CCl4 molecules.The survival of atomic caesium in the rotor chamber in the presence of CC14 vapour seems unlikely ; without any carbon tetrachloride in the rotor chamber the pulses due to caesium reacting to the detector directly from the rotor were only a fifth as large as ion pulses observed in the main experiment. Some experiments were performed in which mercury vapour replaced the carbon tetrachloride ; in this case, elastic collisions only are possible, and the speed of the forward- projected caesium atoms should be 1.2 times that of the incident mercury atoms. In three experiments, the ratio was found to lie between 1.16 and 1.27. DISCUSSION The results summarized above show that it is possible to obtain intense beams of heavy atoms or molecules, moving at known and roughly equal speeds consider- ably greater than those of thermal agitation at ordinary temperatures. It appears possible to study the collisions they make in crossing another stream of atoms or molecules, by observing the times of arrival of the products of such collisions at a suitable detector.Other methods of producing molecular beams for collision studies include the neutralization of positive ions of known specd,3,4 and the selection by rotating shutters of a portion of the velocity-spectrum of a molecular beam emerging from an oven.% 6 The neutralized-ion method is at its best for kinetic energies substanti- ally higher than those obtainable by the present method, while the oven-and-shutter technique is restricted to energies within the thermal range.The rotor method has the disadvantage of being suitable for comparatively heavy molecules only, but the advantage of providing a relatively high intensity ; the flux of neutral molecules through the collision region in the present experiments was the equivalent of several microamperes of singly-charged ions. All three methods seem to deserve further development ; for in spite of their chemical interest, single collisions between neutral atoms and molecules have hitherto been less studied than collisions involving ions or electrons. We should like to acknowledge many stimulating discussions with Dr. D. G. Marshall, who made some of the preliminary experiments leading to those here reported. The rotors were skilfully made by Mr. J. B. Saul. Part of the apparatus was obtained through a Royal Society Government Grant, and one of us (T. H. B.) is indebted to the Department of Scientific and Industrial Research for a mainten- ance award. 1 Marshall, Moon, Robinson and Stringer, J. Sci. Instr., 1948, 167, 478, 2 Moon, J . Appl. Physics, 1953, 4, 97. 3 Horton and Millest, Proc. Roy. SOC. A , 1946, 85, 381. 4 Amdur, Kells and Davcnport, J. Chem. Physics, 1950, 18, 1676. 5 Kofsky and Levinstein, Physic. Rev., 1948, 74, 500. 6 Marple and Levinstein, Physic. Rev., 1950, 79, 223.