Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 354, 17-27 (2000)

Previous Section Next Section Title Page Table of Contents

3. Analysis and results

The flux calibrated spectra of the observed quasars are shown in Fig. 1 and 2. The flux is given in units of 10-15 erg s- 1 cm-2 Å-1. The strongest lines in the spectra are Ly[FORMULA], CIV1549, and CIII]1909. The broad [FORMULA]1400 Å feature which consists of the SiIV1397,1403 and the OIV]1402 multiplet will be refered as SiIV1402 in the following. The strength of the SiIV1402 emission line displays a wide range relative to CIV1549. While HS 1700+64 and UM 667 show strong SiIV1402 emission it is much weaker in 4C25.05 and Q2231-0015 while it is nearly missing in Q 0044-273 (Figs. 1,3).

[FIGURE] Fig. 2. The spectra of the observed QSOs. The flux is given in units of 10-15 erg s-1 cm-2 Å-1. For each QSO the corresponding power law fit [FORMULA] of a nonstellar continuum is displayed.

In addition to these emission lines several important diagnostic lines are visible like OVI1034, OI1304, CII1335, HeII1640, and OIII]1663. The intercombination line of NIII]1750 can be detected in UM196, Q1548+0917, HS1700+64, and Q2231-0015. For Q0103-294 an upper limit of the line strength of NIII]1750 can be estimated at least. Before we measured the integrated emission line flux we had to correct the line profiles for absorption line contamination. Although, one of the selection criteria of the quasar sample was a minor influence of absorption line systems it cannot be totally avoided for quasars at the observed high redshift of z [FORMULA] 3. In the short wavelength regime, [FORMULA] Å, Ly[FORMULA] absorption systems are visible while on the red side of Ly[FORMULA] some metal absorption line systems can be seen (Figs. 1, 2). We restored the contaminated emission line profiles by linear interpolation between adjacent peaks of the observed profile. To illustrate the method the result of this approach is shown in Fig. 3 for the Ly[FORMULA] profile of UM 196 and HS 1700+64. We performed the correction several times independently to the individual spectra to minimize the uncertainties introduced by the interpolation method described above. The corrected individual spectra of the distinct objects were averaged. The method to recover the line profiles by linear interpolation is sensitive to the spectral resolution. Therefore, we rebinned the spectra which we obtained at Calar Alto Observatory/Spain and McDonald Observatory/USA to a lower spectral resolution and compared those with our corrected profiles. The peakness of the rebinned profile core was degraded as expected while the difference in the wings is of the order of less than 5 %.

[FIGURE] Fig. 3. Correction of absorption systems in the emission line profile of Ly[FORMULA] is displayed for the narrow-line quasar UM 196 (top panel) as well as for the broad-line quasar HS 1700+64 (bottom panel). The observed line profiles are plotted as a thin line and the corrected profiles are displayed as bold lines.

For the quasars observed with the VLT UT 1 Antu we obtained spectra of low and high spectral resolution for the Ly[FORMULA] range. The Ly[FORMULA] profile was corrected for absorption features independently. The corrected profiles of the different spectral resolution are in good agreement.

We used the corrected emission-line profiles to measure the line flux. Because the NV1240 and HeII1640 profiles are blended by other strong emission lines we used the CIV1549 emission line as a template profile. Since CIV1549, NV1240, and HeII1640 are high ionization lines (HIL) this approach can be justified. Assuming the line profiles have the same line width and shape the CIV profile we scaled and subtracted it from the emission line profile to measure after transforming both into velocity space. For the transformation of the line profile into velocity space we used the redshifts given in Table 1. This approach does not take into account the contribution of a narrow line component and assumes implicitly that the narrow line contribution of CIV1549 is of the same order for the other broad emission lines which we fitted. Since the broad emission line profiles of the quasars which we studied show no indication of a significant narrow line component, for example in the residua after subtraction of the scaled CIV1549 profile, this influence was neglected.

In Fig. 4 and 5 typical results of the deblending of the Ly[FORMULA], NV1240 and CIV1549, HeII1640, OIII]1663 line profile complexes are shown. This method could be applied to almost all emission lines, solely the Ly[FORMULA] profile differs from the CIV1549 profile. The results of the flux measurements of our quasar sample are given in Table 2. The given uncertainties are estimated from the line fit using the scaled CIV1549 line profile to obtain a reasonable residuum. For the stronger lines like Ly[FORMULA], NV1240, SiIV1400, CIV1549, and CIII]1909 the errors are of the order of 5 - 10 % while it is [FORMULA]20 % or even more for the weaker lines. In addition to the flux measurements a power law fit representing the non-stellar continuum was calculated. The power law [FORMULA] was calculated using those spectral regions which are free of detectable emission lines to fix the continuum intensity. The spectral indices [FORMULA] are given in Table 2.

[FIGURE] Fig. 4. Reconstruction of the Ly[FORMULA], NV1240 line profile complex. The individual profiles are displayed as bold lines and the observed line profile is displayed as thin line.

[FIGURE] Fig. 5. Reconstruction of the CIV1549, HeII1640, OIII]1663 line profile complex transformed to velocity space centered to HeII1640. The individual profiles, as scaled template profiles, are displayed as bold lines, the observed line complex is shown as the upper thin line, and the residuum is displayed as the lower thin line. The flux is given in arbitrary units.


Table 2. Broad emission-line flux measurements for the observed quasar sample. In addition the spectral index [FORMULA] of the power law continuum fit [FORMULA] is given.

We calculated the diagnostic line ratios NV1240/CIV1549 and NV1240/HeII1640 using the flux measurements given in Table 2. Both line ratios (Fig. 6) are in good agreement with the results obtained by Hamann & Ferland (1992, 1993) for quasars at similar redshift. These individual line ratios were used to calculate an average line ratio yielding NV1240/CIV1549 = (0.7[FORMULA]0.3) and NV1240/HeII1640 = (5.9[FORMULA]3.6). The dotted lines (Fig. 6) indicate the line ratios which are expected for the conditions of the BELR gas assuming solar metallicity (Hamann & Ferland 1999). The observed line ratios are obviously larger than those for solar metallicities indicating super-solar abundances. The NV1240/HeII1640 line ratio we derived for the quasar spectra contribute a significant number of new measurements for the redshift range larger than z [FORMULA] 3.

[FIGURE] Fig. 6. NV1240/CIV1549 and NV1240/HeII1640 as a function of redshift. The dotted lines indicate the line ratios which are expected for the conditions of the BELR gas assuming solar metallicity (Hamann & Ferland 1999).

We also calculated the mean NIII]1750/OIII]1663 line ratio for the four quasars we could measure the NIII]1750 emission line flux. The average of this ratios amounts to NIII]1750/OIII]1663 = (0.72[FORMULA]0.26). This rms-value does not take into account the uncertainties of the individual line measurements of OIII]1663 and NIII]1750.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: January 31, 2000