The short-duration afterglow or pink afterglow of nitrogen is superimposed on the well-known long-duration, Lewis–Rayleigh afterglow as a bright flash of milliseconds duration, separated from the discharge by a short delay. The spectrum shows the same band systems as does the discharge, with different relative intensities. The flash of light coincides with flashes of heat and ionization. A comprehensive mechanism, including the various modifications of the short-duration afterglow, is presented. During the delay time, metastable molecules, generated by the ever-present Lewis–Rayleigh afterglow, accumulate until their concentration is sufficient for collisions between two metastables to occur. These, supported by high vibration, generate ions N2+. Their excitation is due to fast electrons or highly vibrating molecules. The periodic occurrence of the pink afterglow is attributed to temperature effects. The blue and green afterglows are explained by modifications of the same idea.

These recollections were written at the request of the Project on the History of Recent Physics of the American Institute of Physics.

Cross sections are measured for excitation of the (0, 0) first negative band by protons in the range 5000 to 25,000 ev and by electrons in the range 45 to 25,000 ev. Upper limits are given for the very weak proton excitation of the first and second positive systems. To find the cross sections for the primary collisions and to avoid secondary effects, measurements are extrapolated to zero pressure. Not all intensities emitted are proportional to the pressure. The processes responsible for these deviations, excitation by neutral atoms and collisional deactivation, are investigated.

We have studied emission of the four forbidden lines of atomic oxygen, λλ 6364, 6300, 5577, and 2972 Å from a discharge through flowing tank oxygen at pressures of a few torricelli, in a tube of 8 mm i.d., with 40–120 mA of current flowing. We found that under these discharge conditions none of the lines were perceptibly broadened or shifted (to within 0.2 Å), and that none showed any response to changes in conditions that would suggest collisional stimulation of emission. Hence we conclude that the emission of these lines is governed by their true radiative transition probabilities. Typical intensities were, for the four lines (in units of photons per cubic centimeter·second): 3.3×109, 1.0×1010, 4.0×1010, and 1.8×109, respectively. The ratio of intensities of 5577 and 2972 Å (which have the common upper level 1S) has a constant value of 22±2 within the range of conditions in which we can observe these lines, as compared with the theoretical value of 16 computed by Garstang. We discuss, in terms of molecular selection rules, the collisional stimulation of the line 5577 Å that is observed by others at high pressures.

We have measured the probability of transition of the forbidden Vegard—Kaplan system (A3Σ—X1Σ) in N2, working from the definition of this probability as the ratio of number of photons emitted per second to number of excited molecules, both measured in the same discharge. We measure the absolute intensity of the 0,6 Vegard—Kaplan band as emitted from the discharge. Under identical conditions we find the population of molecules in the A3Σ level by measuring the strength of absorption of the 1,0 first positive band. By calculation we relate the observed loss (which is a sum of absorption by all rotational levels) to the absolute population of the excited level, in terms of the probability of transition of the first positive system. Our measurements of absorption and emission give a value of (1.6±0.4) × 106 for the ratio of the transition probability of the first positive system to that of the Vegard—Kaplan system. From recent shock-tube measurements of the former transition probability we obtain a value of 2.0(±0.9) sec for the lifetime of the A3Σ state of N2.

Various origins are considered for the atmospheric sodium observed in the twilight and nightglow above an altitude of 70 km. The transport of sea salt to these altitudes in sufficient quantities seems impossible except in the unlikely case of complete atomization of the sodium at low altitudes (30 km). Other terrestrial sources seem even more unlikely. The sodium has not accreted from the sun, as shown by a consideration of the amount of neon now present in the atmosphere. Meteoric influx does provide sufficient quantities of atomic sodium at altitudes of 70 km and greater, and is probably the source of the airglow sodium. The influx of micrometeoritic material is capable of providing even more sodium than do the meteors, but there is great uncertainty as to the fraction of this material which vaporizes. Other metals which have been observed in the high atmosphere are Ca, Li, and Mg. However, no conclusions with regard to the origin of these constituents can be drawn from the presently available evidence.

To measure small concentrations of molecules, particularly in a metastable state, we have developed a form of double-beam absorption spectrometer. This instrument allows one to use as an absorber a tube in which a dc discharge creates unstable species. The design provides that light emitted from this discharge shall not interfere with the absorption measurement. We use as a background source a second discharge tube emitting the band system whose absorption we wish to observe. We measure the absorption of a single vibrational band of known probability of transition. A detailed analysis of the relative intensities of all the lines constituting the band is then necessary in order to extract the population of the metastable level.

When an electron beam of more than 10·5 eV energy traverses low pressure nitrogen the beam is surrounded by a diffuse glow emitting the first positive bands (1). This light is interpreted as an afterglow. It is attributed to metastable molecules, a1Π, which diffuse out of the space occupied by the beam, are thrown by collisions with normal molecules into the unstable state B3Π and then emit the first positive bands. The time required for diffusion is of the order 5 × 10−5 sec. This afterglow is compared with four other types of nitrogen afterglow.

Measurements were made on pure ${\mathrm{N}}_{2}$, using standard microwave cavity techniques, of the coefficient of electron recombination, $\ensuremath{\alpha}$. Pressures ranged from 3 to 30 mm Hg. Spectrograms were obtained of the afterglow. Higher vibrational levels of the second positive bands are enhanced relative to those obtained from electronic excitation. Goldstein-Kaplan bands, the green bands of Gaydon, and the Herman bands in the range 7000-9000 A were also relatively much stronger. The (0,0) first negative band was observed 1500 microseconds after the end of the discharge. Light intensity measurements by means of a photomultiplier tube showed a peculiar behavior of the relative intensities of the first positive and second positive bands. The number of photons emitted was large as compared to the number of electron-ion pairs recombining.In the afterglow of oxygen the only light observed was the (0,0) atmospheric band and weak radiation in the region 3500 A to 4500 A, which is thought to be a continuum. The continuum decreases in intensity as though it were the result of a two-body process.

For the observation of weak absorption lines a background is desired that shows as small fluctuations as possible. The light from a low pressure discharge serves this purpose under the following conditions: a dc discharge, viewed end on, using a plane cathode mounted in a glass tube of larger diameter; the pressure that provides the highest conductivity of the discharge or a somewhat higher rather than lower pressure; a power supply of high voltage combined with a large series resistance. Under favorable conditions the light from the discharge produces a photoelectric current whose fluctuations are only about twice the inevitable shot noise.

By a consistent argument, based on the characteristic curves of photographic plates, the optimum conditions for the following spectrographic problems are derived: weak emission line without background; weak absorption line; weak emission line on a given continuous background; prefogging; measurement of relative intensities. The effect of the grain on the resolving power is minimized by averaging over the length of the line. This technique gives to the very sensitive plate (Kodak type I-O), in spite of its coarse grain, the same high resolving power as that of a very fine grain plate which requires a 90 times longer exposure.

The process by which an excited Kr atom in an impact of the second kind dissociates an N2 molecule is investigated by comparison with related processes. (1) The production of free N atoms is not due to an atom transfer reaction. (2) It is not due to the ``resonance'' transfer of electronic energy of Kr′ directly into vibrational energy of N2 leading to dissociation. (3) It is due to the transfer of electronic energy of Kr′ into electronic energy of N2. The simplest interpretation is the production of a repulsive level of N2 which immediately dissociates.

The theory of the "Lewis-Rayleigh afterglow" of nitrogen, originally given by Cario and Kaplan, is modified by taking into account recent work regarding resonance in collision processes and the value of the energy of dissociation of ${\mathrm{N}}_{2}$, accepted in this paper as 9.76 ev. Kaplan's "auroral afterglow," in which the spectrum of ${\mathrm{N}}_{2}^{+}$ is observed, is tentatively explained by ionization caused by collisions of two metastable molecules. The results are applied to the airglow.

The limitations are investigated under which "resonance" provides certain collision processes with preferred probability. In impacts of the second kind, resonance strongly affects the transfer of electronic into electronic energy or into a small amount of vibrational energy, possibly the transfer of vibration into vibration, certainly not the transfer of vibration into rotation. Resonance is not effective in the transfer of energy of atomic recombination into electronic energy and, in general, resonance is not effective in processes in which nuclei change positions by appreciable amounts.

Das Spektrum des sonnenbestrahlten Nordlichtes besteht im wesentlichen aus den blauen Banden des N2+-Ions und den roten, verbotenen Linien des O-Atoms. Während nachBates die Banden des N2+ durchzwei aufeinanderfolgende Vorgänge (Stoßionisation und Fluoreszenzanregung durch Sonnenlicht) hervorgerufen werden, müssen wir für die Linien des O Anregung ineinem Vorgang, vermutlich Elektronenstoß, annehmen. Diese Unterscheidung läßt erwarten, daß der blaue Bestandteil eine viel längere Lebensdauer hat als der rote und deshalb viel stärker dem Abtreiben durch Winde ausgesetzt ist.

In the behavior of the thermal hydrogen-oxygen reaction there is an apparent discrepancy. The second explosion of the thermal reaction occurs over a wide range of low temperatures down to 350°C, even though the rate of steady thermal reaction is very nearly zero. The authors used a criterion stated by Lewis and von Elbe to show that the same volume chain which explains the reaction at higher temperatures applies to the low temperature range. The inhibition period within the low pressure explosion region and the branching of the chains are briefly discussed.