Vibrationally excited IO( νννν '← ← ←νννν ← ") with νννν ">0

In the spectral window from 467nm to 600nm (the red end of the observational window) the contributions of vibrationally excited IO(ν'←ν"), ν">0, of OIO and of I₂ could be separated with a good quality and low errors wherever the spectrum has significant absorption.

Collinearities with ground state IO were avoided, as the IO(0←0) at 465.5nm was excluded from the spectral window. By limiting the spectral window to 480 to 600nm this could be improved even more, because then IO(ν'←1) with its different behaviour was excluded.

Especially the low pressure data sets containing the largest concentrations of excited IO yielded clear spectra. For the bands falling between those of ground state IO, the separation proved to be more difficult, as there the similarity of temporal behaviour of ground state IO in general, that of IO(2←0) and that of excited IO produce collinearities and imperfect extraction of individual spectra. On the other hand the temporal behaviours were sufficiently different to inhibit a reduction of free parameters. In this interval the noise in the extracted spectra is larger and negative traces of ground state IO remain. Below 440nm again the absorption of absorber "Z" needs to be taken into account. But in a number of data sets in the separations with PCA/ICA as well as in the extraction with multivariate regression a further absorption, labelled "X", proved to be significant and was in many fits extracted together with excited IO as a single spectrum with acceptable residuals. The new absorption displays a smooth continuum located in the region of the ground state IO continuum and shows a temporal behaviour similar but not completely the same as that of excited IO.

The temporal resolution of our set-up is defined on one hand by the shifting time per row of the CCD chip and on the other hand by the characteristic function in time (defined by the masking of the chip and the vertical slit width of the mask). The characteristic function has a

half width of 5 to 7 pixels in different regions of the chip. The variations are caused by mechanical imperfections of the laboratory-made masking of the chip. The difference between excited IO and the found absorption is visible only in a time interval of 10 to 12 pixels in time – roughly twice the width of the instrument's characteristic function in time – and very close to the flash. The large remainder of both curves is nearly identical. Therefore it is possible that the observed continuum belongs to excited IO. But on the other hand it can not be excluded that it originates from a further not yet identified species. This will be decided after further analysis of spectra, see below (Chapter 8). The full spectrum obtained by joining the three different sections is shown in the middle diagram of Fig. 7.4. The top diagram shows an original mixed spectrum for comparison and orientation. In the bottom graph the obtained relative uncertainty is plotted.

Uncertainty and reproducibility

The reproducibility of the spectrum was highest in the range from 480 to 600nm. The shape of the bands and zero absorption in the troughs was always reproduced. Depending on the mixture changes in relative height between groups of bands from different ν" occurred. When the window was extended down to 467nm and below, the different behaviour of IO(ν'←1) and IO(0←0) reduced the reproducibility in that traces of ground state IO appeared in the spectrum or a continuous background disturbed the spectrum. In the continuum in spite of the low signal to noise the reproducibility was good. Based on the error estimate the uncertainty of the spectrum is approximately 2 to 3% in the peaks. In the continuum range it is of the order of 5 to 10%. Between the bands the absolute error is of the order of 2 to 5⋅10^-4 in units of optical density.

Wavelength calibration

The wavelength calibration for vibrationally excited IO is the same one obtained from the line source measurements, as described for ground state IO.

Assignment of transitions

In 1958 [Durie and Ramsay 1958] recorded bands of IO in absorption, but in those measurements apart from the absorption transitions from ground state only the IO(2←1) was visible. Durie et al. [1960] carefully examined the emission spectrum of IO using a 21ft grating spectrometer in second or higher order and observed a large number of transitions

300 350 400 450 500 550 600 -0,005

0,000 0,005 0,010 0,015 0,020 0,025 0,030 0,035

300 350 400 450 500 550 600

0,01 0,1 1 10

300 350 400 450 500 550 600

-0,2 0,0 0,2 0,4

IO(

'

") with

'>0

optical density

vacuum wavelength (nm) 2←←←←1

rel. error optical density

IO

I

OIO

higher oxides

Figure 7.4 The full spectrum of vibrationally excited IO obtained by joining the three sections from the three extraction intervals is shown in the middle diagram. In an original mixture its spectrum is covered by spectra of ground state IO, higher oxides, I2 and OIO (top diagram). The (2←1) band again shows an apparently reduced height. The relative uncertainty of the extracted spectrum in the continuum range is of the order of 5 to 10% while in the peaks it decreases to 2 to 3% (bottom diagram, open circles). Between the bands the absolute error is of the order of 2 to 5⋅10^-4 in units of optical density. The assignment of the continuous absorption to excited IO is not unequivocal as the width of its temporal behaviour is of the order of the width of the detector's characteristic function in time. It could also be caused by a further yet unidentified species.

They performed a thorough rotational and vibrational analysis, which enabled determination of the band origins – i.e. that transition within a vibrational band, where J'=0 and J"=0 - of both ground state and vibrationally excited IO. From the band origins they determined vibrational constants for IO, with residuals generally better than 0.005nm. Only overlapped bands were reproduced with larger residuals of 0.02 to 0.03nm.

In our low resolution measurements for the first time a large number of vibrationally excited bands are observed in absorption. In our extracted absorption spectrum 22 bands of excited IO out of 34 observed by Durie et al. [1960] in emission are present. Fig. 7.5 shows our spectrum of transitions from vibrationally excited levels with the assignments according to Durie et al.

As our measurements were aimed at broad and simultaneous spectral coverage, high resolution was not the point of our work. Therefore our spectra can not compete in resolution with their work. Given the limited resolution of 1.2nm FWHM and 0.35nm per pixel of our measurements, our band heads are in reasonable agreement with their accurate band origins.

Examination of band height of individual progressions shows that in the IO(ν'←1) similar as in the ν"=0 series again the transition leading to the A²Π3/2, ν'=2 appears to be reduced in comparison to the neighbouring bands of the same series (see Fig. 7.4 and compare also Fig.

3.1). This indicates that the observed anomaly is linked to the A²Π3/2, ν'=2 state. A closer examination of temporal behaviour was impeded by collinearities and by the for our set-up too short time scale of formation and consumption of these excited species.

400 450 500 550

-0,005 0,000 0,005 0,010 0,015 0,020 0,025 0,030 0,035

1←←←←3 0←←←←3 1←←←←4

3←←←←4 2←←←←4 0←←←←4

2←←←←5 1←←←←5 0←←←←5

2←←←←2 1←←←←2 0←←←←2

0←←←←1

1←←←←1

2←←←←1

3←←←←1 1←←←←6 1←←←←70←←←←6

optical density

vacuum wavelength (nm) 3←←←←7 4←←←←1

Figure 7.5 The observed bands are identified using the assignations made by the high resolution study of Durie et al. [1960], who observed the emission spectrum of IO using a 21ft spectrograph in orders 2 or higher. While the IO(ν'←ν") for ν"=1 shows the first four transitions clearly, it is remarkable that for higher progressions dominantly the (0←ν") occurs in the spectrum. As it will turn out later (Chapter 8), this directly reflects the relative magnitude of Franck-Condon factors.

The second interesting feature of the observed hot bands of IO is the strong and regular presence of the IO(0←ν") transitions. Starting with the barely visible (0←6) the bands increase steadily until the (0←2), before decreasing in the (0←1) again.