3. Sample preparation and spectral measurements
In the previous study (Nuth et al. 1985) the P&B "gray rocky" samples were ground with an agate mortar and pestle, diluted with granular polyethylene, heated to about (380-390 K) to form a pellet, and the transmission spectrum measured at room temperature from 80-700 (125-14 microns). In the present study, two different sample preparation methods were used which gave very similar spectra. In both of these methods, the sample was ground in an agate mortar and pestle, diluted with KBr to a concentration of 10-15% , and the diffuse reflectance spectrum was measured from 4000 to 400 (2.5-25 microns). The difference between the two types of measurements was in the use of an Harrick environmental chamber for taking measurements at 600 K in vacuo as opposed to measuring the spectrum at room temperature in air. In the experiments using the environmental chamber, it was evacuated using a mechanical vacuum pump to the milliTorr level before and during heating and during spectral measurement. Typically the spectrum was measured both at room temperature (300 K) and then again after the temperature stabilized at 600 K. In a second series of measurements diffuse reflectance spectra were measured in air at room temperature and then again after the same sample had been heated in air at (380 K) for periods of up to one hour and then cooled back to room temperature. Aside from a transient feature seen briefly in the environmental chamber spectra upon reaching 600 K, the spectra obtained by both these techniques were consistent with each other. In the overlap region from 400-700 (25-14 microns) both sets of spectra were completely consistent with the earlier spectrum of Nuth et al. (1985). We attribute the transient feature at 600 K to traces of released by pumping from the sample and reabsorbed onto the larger surface of the KBr dilutent. could easily form by reaction of the with an atmospheric water contaminant picked up during sample preparation. This reaction would result in a small amount of silica contaminant in the sample.
Because Begemann et al. (1996) had stressed the very high reaction rate of with humid air and because they had also demonstrated that both crystalline and glassy yield similar (though not identical) spectra, we tried a new set of experiments designed to address the issue of reactive contamination. Strem Chemicals synthesized "cotton-like lumps" guaranteed to be more than 99.5% pure : they shipped this material to us double bagged in dry nitrogen. A sample was removed in a nitrogen filled glove bag, ground with KBr and measured in diffuse reflectance. A second, larger sample was removed for XRD measurement in air. The remaining material was stored in a desiccator in the original bottle after flushing with dry . The XRD pattern is completely consistent with crystalline as shown by comparison with data for orthorhombic from the Joint Center for Powder Diffraction Studies (JCPDS 1988) (Fig. 1). The FTIR spectrum of this sample contains significant absorbance from 8-13 microns as shown in Fig. 2a.
The sample used for XRD was left exposed to air for 1.25 hours in the hood after we obtained the XRD pattern; it gave the spectrum shown in Fig. 2b. Subtraction of the spectrum in Fig. 2b from that in Fig. 2a yields a spectrum shown in Fig. 2c that is similar to that reported by Begemann et al. (1996) for . We remeasured the diffuse reflectance spectrum of a sample of removed from the sample bottle after a months storage in dry and obtained a result virtually identical to Fig. 2a. Finally we placed a sample of in the FTIR and measured its diffuse reflectance spectrum each hour for more than two days without observing any significant change due to exposure to the air in our FTIR. Note that this air is dried to a dew point of by a Balston Air Drying Unit. These studies demonstrated to our satisfaction that is reasonably stable in dry air and that the sharp features observed at 20, 17 and 12 microns are extremely sensitive to humidity.
Fig. 3a-3d are the spectra of various samples measured in dry air at 300 K. Fig. 3e-3h are spectra of samples measured in vacuo at 600 K. All of the spectra in Fig. 3 share some common features with one another as well as some significant differences from the spectra of Begemann et al. (1996) and the spectrum in Fig. 2a. The most significant difference between the spectra in Fig. 3 and that in Fig. 2a is the lack of sharp features at 12 and 17 microns as well as a shift and broadening of the peak absorption from to microns in the samples shown in Fig. 3. However, the similarities in the spectra are extremely significant. All of the spectra in Fig. 3 show features near 7.5 (shoulder), 8.5, 9.5, 12.5 and 21- 22 microns. All spectra taken at room temperature also show a feature at 10.7 microns that disappears on heating to 600 K. The spectrum in Fig. 2a shows a shoulder near 8.5 microns as well as distinct features near 9.3, 10.8, 12.2 and 22.5 microns, although the intensities of the peaks differ somewhat from the pattern displayed by the spectra in Fig. 3.
We also observed two additional features that appear to be due to impurities. A peak near 15 microns which disappears on heating is due to a impurity present in KBr. A feature at 23.9 microns (not shown) appeared (briefly) only in those samples whose spectra were run in the environmental chamber. This feature disappeared on long heating. It was not present in any samples which were run in air whether heated or not. It is suspected that this feature may be due to a volatile impurity, possibly , which is initially trapped in the matrix but which on evacuation and heating of the environmental chamber is released and readsorbs on the much larger surface area of the KBr dilutent before being pumped away.
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998