Astron. Astrophys. 329, 291-314 (1998)

8. Beryllium iso-electronic sequence

The EUV spectra of the Beryllium-like ions is dominated by the 2s^{2 1} S₀ - 2s2p ¹ P₁ transition which can be seen in the SERTS-89 observations from consecutive elements sodium through to sulphur. Of these elements, it is only for Mg IX and Si XI that we find lines other than this, namely the 2s^{2 3}P - 2s2p ³P transitions. For O V, weak 2p - 3s lines are identified.

8.1. O V

The two O V lines in the Thomas & Neupert catalogue are both very weak identifications: 215.29 Å is a weak feature seen in the wing of first order Mg VIII 430.45 Å, while the 248.46 Å occurs close to the low wavelength end of the spectrometer bandpass, where the sensitivity is lowest.

The 215.29 Å line seems unlikely to be a O V feature as we would then have expected the O V 220.35 Å line to have appeared in second order at around 440.70 Å, as it has an intensity at least as strong as the 215.29 Å line. However, nothing is seen there despite it being a fairly "blank" part of the spectrum.

We show the 248.46/220.35 ratio in Table 4 to give an indication of the expected intensity of the 220.35 Å line. The error bars on the theoretical ratio are due principally to temperature sensitivity which is quite marked for these two lines despite their closeness in wavelength.

8.2. Mg IX

A comparison of SERTS-89 Mg IX intensities with theory was provided by Keenan et al. (1994a). We note that they used older RMATRX calculations (Keenan 1988, Berrington 1985) whereas in the CHIANTI database we use the more recent Distorted Wave calculations of Zhang & Sampson (1992).

For the six identified lines we find three branching ratios which are given in Table  3. The 448.28/441.22 ratio is discrepant by almost a factor of two with the 448.28 Å line being weaker than expected. Keenan et al. find a similar result, and suggest that this is an instrumental effect due to the 448.28 Å line falling close to the high wavelength end of the SERTS-89 bandpass.

Table 3. Branching ratios for the Beryllium-like ions.

There are two regions of density sensitivity in Mg IX: first the 2s2p ³ P₀ and ³ P₂ levels gain significant populations (relative to the ground ¹ S₀ level) at 10 cm^-3 and then the 2s2p ³ P₁ level gains a significant population for densities greater than 10 cm^-3. Between 10 cm^-3 and 10 cm^-3 the Mg IX ratios are density insensitive and the ratios reported in Table 4 apply only to this density regime for these lines.

Table 4. Insensitive ratios for the Be-like ions.

The 441.22 Å and 448.28 Å lines give density insensitive ratios relative to 443.96 Å and we give both ratios in Table  4. They potentially allow a determination of which of the 441.22 Å and 448.28 Å lines is in error, but both ratios are consistent within the error bars on the data.

The low value of the observed 443.96/368.06 ratio means that it falls within the 10 cm^-3 density sensitive regime and so we give the density that it indicates in Table 5. Similarly, although the 441.22/443.96 insensitive ratio agrees with theory within the error bars on the data, the fact that the observed value lies above the theoretical value means that the same ratio can be used to derive a density. In both cases the density is somewhat low and not consistent with other ions formed at the same temperature.

Table 5. Density sensitive line ratios for the Be-like ions.

In their analysis, Keenan et al. (1994a) assume a density of 10 cm^-3 and compare the observed line intensities with theory, finding good agreement apart from with the aforementioned 448.28 Å line. We note that although their theoretical value for the 443.96/368.06 ratio is smaller than ours (0.026 compared to 0.030), it remains outside the error bars of the observed data. The reason for their smaller predicted ratio stems mainly from their larger value of the 368.06 Å thermally-averaged collision strength (1.59 compared to our 1.20).

The disagreement of the Mg IX 443.96/368.06 ratio with theory may reflect an error in the intensity calibration curve of SERTS-89, as discussed later in Sect. 15.1.

8.3. Si XI

The excellent spectral resolution of the SERTS instrument ensures that the Si XI 303.32 Å line is well separated from the far stronger He II line at 303.78 Å, which has not been possible with previous imaging spectrometers. Only three of the five ³P - ³P transitions are identified in SERTS-89, and there is relative density sensitivity between the ³P lines and the strong 303.32 Å line in the density range 10⁷ -10¹⁰ cm^-3. Care has to be taken below densities of around 10 cm^-3 as photo-excitation by a background radiation field can play an important role in de-populating the 2s2p P level, significantly affecting the level balance of the ion. This process is not accounted for in the current version (1.0) of CHIANTI.

The two ³P - ³P transitions not identified by Thomas & Neupert are predicted by CHIANTI to be at 358.65 Å and 364.50 Å and are both strong enough to have been observed, as witnessed by the branching ratios presented in Table  3. Both lines are however blended with other stronger lines and so we use the branching ratios to predict Si XI contributions to each blend.

The CHIANTI 358.65 Å line is blended with the reported Fe XI 358.67 Å line. As discussed in Sect. 13.3, the latter line at 73 erg cm^-2 s^-1 sr^-1 is a factor 3 too strong, consistent with the presence of a blend; however, the Si XI component only provides around 14 erg cm^-2 s^-1 sr^-1. Other components are suggested to be from Ne IV and Fe XIV.

The line expected at 364.50 Å is blended with the stronger Fe XII 364.47 Å line. From the branching ratios we can estimate a Si XI contribution of 17 erg cm^-2 s^-1 sr^-1 to the observed total intensity of 233 erg cm^-2 s^-1 sr^-1 for this feature.

The observed 371.50/361.41 ratio is not in agreement with theory and although there is an Fe X line predicted by CHIANTI at 361.40 Å it is not strong enough to account for the discrepancy. Of interest is the fact that observations for the corresponding transitions in Mg IX (448.28/441.22) were also inconsistent with the theoretical branching ratio and by a similar amount, the 371.50 Å and 448.28 Å lines being both around a factor of two too low. This suggests that there may be a problem with the transition probabilities for these two transitions.

As the 2p P and P levels do not come into Boltzmann equilibrium with the ground level until densities 10 cm^-3, we can use the 365.43/303.32 ratio to estimate a density (given in Table 5) that is consistent with other elements formed at the same temperature.

Above 10 cm^-3, the 361.41/365.43 and 371.50/365.43 ratios become density insensitive with values shown in Table  4. Interestingly, it is the 371.50/365.43 ratio rather than the 361.41/365.43 ratio which agrees better with theory, which suggests that the 361.41 Å line may be blended.

8.4. Summary

The relative density sensitivity of the beryllium-like ion lines tends to be small over typical coronal densities of 10 - 10 cm^-3 and so the line ratios are not useful as density diagnostics, apart from the Si XI 365.43/303.32 ratio. The main features of the Be-like lines are as follows.

• The S - P transition does not show agreement with the P - P transitions for Mg IX unless a small density of around 10 cm^-3 is assumed. The same ratios for Si XI can be used to give a density consistent with other ions formed at similar temperature, and so we suggest that it is only for Mg IX that there is a real problem. One possibility is that the SERTS-89 calibration may require some adjustment above 430 Å.
• Observed ratios of ( P - P)/(P - P) lines for both Mg IX and Si XI are around a factor of two too low compared with predicted branching ratios, suggesting that presently accepted transition probabilities for both ions are incorrect. On the other hand, such a large error is difficult to explain in terms of atomic physics, and so an alternative suggestion is that the Mg IX 448.28 Å line is under-estimated due to optical vignetting since it is very close to the high wavelength end of the SERTS-89 bandpass, whereas the Si XI 361.41 Å line may be affected by blends.

© European Southern Observatory (ESO) 1998

Online publication: November 24, 1997
helpdesk.link@springer.de