Dynamic Spectra of Small-Mass Meteors

Abstract. We present dynamic (22 frames per second) observations of optical spectra of small-mass (2–200 mg) meteors observed with an EMCCD imager equipped with a diffraction grating. This observational campaign occurred at Arecibo, Puerto Rico during May 2012, resulting in eight hours of clear data over four nights. We detected 22 meteors with this setup and their spectra showed varying compositions, including evidence of Na, Mg, Fe, and Ca. Spectral lines persisting over multiple frames (up to 23 frames) with sufficient signal, showed evidence for differential ablation. Brighter, more massive meteors had stronger and varied spectral signals, which showed that the temporal and spectral resolution of the faintest meteors approached the noise level of the camera system. Optical and spectral detections of these small-mass meteors provide a greater understanding of the composition of the milligram-sized population of meteors.


. An example raw image of a meteor and its spectra (left) that is uncalibrated and unflattened. An example stellar spectra (right) for a type F0 star is shown. This star was used for calibrating the wavelength sensitivity of the combined imager-diffraction grating system.

Stellar Calibration
The first step in the imager calibration was to use reference stars in the images to convert the intensities from the imager (in counts) into stellar magnitudes. This was accomplished by creating a linear fit between the magnitudes of known reference stars and the logarithm of the intensities (in counts) from the imager. Figure 2 shows this relationship for the four brightest stars in the image, where clear intensities were able to be derived from the zeroth-order spectra (un-diffracted images). For 95 these data, we found a linear fit, with an r 2 value of 0.972 using the four brightest stars in an image. This linear fit was then used to convert the imager counts corresponding to the meteors into stellar magnitude values which were then used to calculate the meteor masses, using the kinetic energy and luminosity relationship found in (McKinley, 1961). For the 22 meteors detected, we can place lower bounds on the mass, because we assumed that the only measured component of the velocity (the horizontal component) was the total speed, knowing the the total speed must have been higher. In addition, without radar returns or 100 another imager for triangulation, we simply assumed that the altitude of all the meteors was 95 km, which will also contribute to the uncertainties in the mass estimates.
In order to characterize the relative intensities of the observed emission lines in the meteor spectra, we need to quantify the imager response with the diffraction grating as a function of wavelength (the spectral response), which was not done in the lab at the time. Due to quantum efficiency effects of the CCD detector itself and the transmission characteristics of all the lenses, 105 there will be a spectral dependence on the measured intensity of the meteor spectra that needs to be normalized out. To do this, a star of a known spectral type (F0 in this case), was examined. Figure 3 shows the measured spectra from an F0 type star, fit to a Planck curve with a temperature of 7350 K. This was then used to create a wavelength normalization that was applied to the meteor spectra, such that the relative intensities of the observed emission lines were of physical significance and not just instrumentally created. Atmospheric transmittance corrections were also applied to each spectrum that affected the shorter 110 wavelengths (350 -400 nm).

Observations
There were a total of 22 meteors observed, where spectra were discernible. Table 1 lists a summary of the main characteristics of these 22 meteors, including the date, time in UT, the horizontal velocity in km/s, the azimuthal direction of origin in degrees where East = 0 and North = 90 , the maximum brightness in visual magnitude and the derived mass in milligrams. The 115 maximum brightnesses ranged between 3.2 and 0.4 magnitude and the masses ranged between 2.4 and 153.1 mg. Table 1, strong spectral lines were visible for Ca, Fe, Mg, and Na in some meteors. Lines were observed at 518 nm and 553 nm for Mg (Savage and Boitnott, 1971), and in all cases where Mg was present, both spectral lines were visible.

As shown in
In addition, strong Na lines were only visible in the brightest meteors that also showed evidence for the other three observed elements. This analysis was done by observing the full evolution of each spectrum, and peaks were made clear by summing 120 over multiple frames. Figures 4 and 5 show the spectrum of meteor 4. This meteor was visible for 11 frames, or 0.5 seconds, and higher wavelengths of the spectrum were not visible for the first half of the meteor's path across the detector. The Fe line is not visible until more than halfway through the meteor's path, while Mg is seen as an excited line throughout the spectrum. Although part of the spectrum is lost due to the limited field of view, the 518 line of Mg is seen in both the beginning and end of the spectrum.