A&A 404, 997-1009 (2003)
T. Iijima1 - H. H. Esenoglu2,3
1 - Astronomical Observatory of Padova, Asiago Section, Osservatorio Astrofisico, 36012 Asiago (Vi), Italy
2 - Astronomy and Space Science Department of Istanbul University, University Observatory, 34452, University, Istanbul, Turkey
3 - Istanbul University Observatory Research and Application Center, 34452, University, Istanbul, Turkey
Received 28 January 2003 / Accepted 21 March 2003
Spectral evolution of the fast nova V1494 Aql was monitored soon after its discovery in December 1999 to September 2000. The first spectra showed prominent emission lines of H I and Fe II, while He I was seen in absorption. The radial velocities of the absorption components of H I, He I and N II rapidly increased (in the negative sense) during the early decline stage, while those of Fe II remained nearly constant. When a new spectrum was taken on February 6, 2000 after the seasonal interruption, this nova was in the transition stage. The spectra in the transition stage showed emission lines of H I, He I, He II, N II, N III, Si II, [N II], [O I], [O III], [Fe II], [Fe VI], [Ca V] etc., hence the emission lines of Fe II had disappeared. A quasi-periodic oscillation of luminosity with a time scale of about days and a mean amplitude of about 1.2 mag in V band was seen from February to the middle of April 2000. The emission lines of He II and [Ca V] disappeared around a light maximum of the oscillation, while the emission lines of N II and N III strengthened. At the same time high velocity (-2900 and +2830 km s-1) broad emission wings of H I lines appeared, which suggest an ejection of high velocity jets. The excitation state increased throughout the nebular stage. The last spectra taken in September 2000 showed highly excited emission lines up to [Fe VII] and [Fe X] 6374.5.
The interstellar extinction is estimated as from the equivalent widths of the interstellar absorption components of Na I D1 and D2. Using this result, the distance to the nova is estimated as kpc. The mass and the helium abundance of the ejecta are estimated as 6.2 1.4 10 and N(He)/N(H) = 0.13 0.01, respectively. The electron density of the ejecta decreased as during the nebular stage, where t is time from light maximum. This low decline rate suggests that the ejecta had a ring like shape as well a large mass loss which may have continued throughout the nebular stage.
Key words: stars: individual: V1494 Aql - novae, cataclysmic variables - ISM: general
V1494 Aql (Nova Aql 1999-II) was discovered by Pereira (1999) on December 1, 1999 as a star of . The luminosity and time of light maximum were estimated as on December 3.4 UT, 1999, JD 2 451 515.9, by Kiss & Thomson (2000). Early spectral evolution was reported by Kiss & Thomson (2000) and Anupama et al. (2001). Kiss & Thomson (2000) obtained the decline rates of the luminosity days and days, and estimated the absolute magnitude at maximum . They obtained a distance to the nova D = 3.6 0.3 kpc. This distance, however, should be an upper limit because the effect of the interstellar extinction was not taken into account. Kawabata et al. (2001) carried out some spectro-polarimetric observations and suggested that the ejecta were not spherically symmetric. Many coronal emission lines were detected by infrared spectroscopy carried out in July 2000 (Venturini et al. 2000) and July 2001 (Rudy et al. 2001).
The monitoring of spectral evolution of this nova in our observatory started soon after the announcement of discovery. The first spectra were taken on December 5, 1999, that is 2.3 days after maximum, then high and medium dispersion spectra were taken until September 16, 2000. In this paper, we report the spectral evolution of this nova in the early decline, transition, and nebular stages. The physical properties of the ejecta are estimated using the intensities of the emission lines in the nebular stage. The high velocity jets observed in the transition stage are also reported.
High dispersion spectra covering the range 4350-6900 Å with a resolution of were taken with an Echelle spectrograph mounted on the 182 cm telescope at the Mount Ekar station of the Astronomical Observatory of Padova. Medium dispersion spectra were taken with a Boller & Chivens grating spectrograph mounted on the 122 cm telescope of the Astrophysical Observatory of Asiago of the University of Padova. The resolution is with a grating of 600 lines mm-1. The reductions of the spectra were carried out using the standard tasks of the NOAO IRAF package at Asiago Observatory. The spectrophotometric calibrations were made using the spectra of standard stars taken in the same nights. HR 7596 (58 Aql) was used for the calibrations of the high dispersion spectra and Kopff 27 for the medium dispersion spectra. A log of the observations is given in Table 1, where UT is the universal time at the start of the exposure and "days'' is number of days from light maximum.
The effect of interstellar extinction was not taken into account in the previous works of this nova, because no interstellar absorption line was detected in their low dispersion spectra (Kiss & Thomson 2000; Anupama et al. 2001). The interstellar absorption components of Na I D1 and D2, however, are clearly seen on our high dispersion spectra (see, e.g. Figs. 11-13). The mean equivalent widths of D1 and D2 are and Å, respectively. Using the double-ratio method of Münch (1968), the column density of sodium atoms is estimated as log(N(Na)L) = atoms cm-2 from these equivalent widths, where L is the length of the line of sight in cm. The amount of the interstellar extinction may be E(B-V) = and , where we adopted the chemical abundance of the interstellar medium, log(N(Na)/N(H)) =-8.4 (Cohen 1975), and the relation between the column density of hydrogen atom and the interstellar extinction, N(H)L/E(B-V) atoms cm (Jenkins & Savage 1974). The observed spectra are, therefore, dereddened assuming E(B-V) =0.6.
If we add the effect of the interstellar extinction, , to the result of Kiss & Thomson (2000), the distance to the nova should be kpc.
|Figure 1: Light curve of V1494 Aql collected from VSNET (http://www.kusastro.kyoto-u.ac.jp/vsnet/index.html).|
|Open with DEXTER|
Figure 1 shows the light curve of V1494 Aql collected from VSNET. This nova showed a fast fading in December 1999. Our first spectra were taken on December 5 with the Boller & Chivens spectrograph, then successive observations in the early decline stage were made on December 6 and 17.
Figures 2-4 show tracings of our first spectra, where the unit of the ordinate is 10-10 erg cm-2 s-1 Å-1. Most prominent emission lines are of H I and Fe II, whereas He I and some N II lines are seen as absorptions. The weak emission line at 6165.4 Å could be a blend of O I 6157.0 and N II 6167.8 (Fig. 4). The emission components of H, H and H show single peak profiles, while H and some Fe II lines show saddle shaped profiles. Some H I and Fe II lines are flanked by P Cygni type absorption components. The absorption components of H and H are very weak and that of H is not detected in our medium dispersion spectra.
|Figure 2: A spectrum of V1494 Aql on Dec. 5, 1999. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 3: A spectrum of V1494 Aql on Dec. 5, 1999. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 4: A spectrum of V1494 Aql on Dec. 5, 1999. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
The spectral features on December 6 were effectively the same as those on December 5, while large changes were noticed on December 17. Figures 5-7 show tracings of the spectra on December 17. A tracing of the spectrum multiplied by ten is over-plotted as a dotted line in Fig. 7 to reveal the weak spectral features. In contrast to the earlier spectra, H and H show profiles with well separated two peaks and a weak central peak. The emission components of H and the higher member of Balmer series are flanked by P Cygni type absorption components. The blue-ward emission peaks of H and H seem to be cut out by the absorption components. A weak trace of [O I] 5577.4 is seen on the spectrum on December 17 (Fig. 6). This line was not seen on any other earlier nor later spectra.
An absorption line is seen at 4608 Å on the spectrum taken on December 17 (Fig. 5). If this line is due to N III 4640.6, its radial velocity should be -2110 km s-1. This velocity, however, is too low with respect to those of the other N II and He I lines (see, Sect. 4.3). This absorption line may be due to a blend of N III 4640.6, N II 4643.1 and O II 4649.1.
|Figure 5: A spectrum of V1494 Aql on Dec. 17, 1999. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 6: A spectrum of V1494 Aql on Dec. 17, 1999. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 7: A spectrum of V1494 Aql on Dec. 17, 1999. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1. The dotted line shows a tracing of the spectrum multiplied by ten.|
|Open with DEXTER|
The changes of the profiles of H also have been reported in previous works (Kiss & Thomson 2000; Anupama et al. 2001). The observations made by Anupama et al. (2001) show that the profile of H largely changed between December 10 and 16. When the profiles of the H I lines changed, their widths significantly increased as shown in Table 2.
A weak emission line is seen at 7105 Å on the spectrum on December 17 (Fig. 7). This line was not observed in any previous novae or nova-like objects, at far as we know. Its laboratory wavelength may be Å. Here, we tentatively identify it as [Ni II] 13F 7102.8 (Moore 1959), but no other line of the same ion, e.g. [Ni II] 14F 5275.8, was detected in the spectra. Further works are necessary for the identification of this line.
Table 2: Intensities of selected emission lines in the early decline stage corrected by E(B-V)= 0.6.
Table 3: Radial velocities of selected absorption lines in the early decline stage.
Intensities after the correction of the interstellar extinction by E(B-V)= 0.6 of some selected emission lines in December 1999 are given in Table 2. The selected lines are relatively free from blending and are not necessarily stronger than the others. The intensity ratio of H/H and the FWHM of H are also presented. The observational errors in the intensities are about 10% and those in the FWHM of H are about 50 km s-1. Intensities affected by larger observational errors are denoted by a colon.
|Figure 8: Intensity ratio of H/H.|
|Open with DEXTER|
Figure 8 shows the variation of the intensity ratio H/H, where the results in the transition and nebular stages are also plotted. The theoretical ratio of H/H in Case B recombination should be found in the range of 2.6-3.0 in various conditions of electron temperature and density (Hummer & Storey 1987). As seen in Fig. 8 and Table 2, however, the ratio H/H of the present nova was lower than the theoretical values on December 5 and 6. The ratio suddenly increased afterwards and became much higher than the theoretical one on December 17. At the same time, the profiles of H and H significantly changed and the width of H increased, as reported in the previous sub-section. The high intensity ratio continued until April 2000, then gradually decreased (see also Sect. 6).
The large ratios of H/H on December 17 and later may not have been due to an ejection of obscure matter, because a deep light minimum such as of DQ Her (e.g. Martin 1989) was not seen in the light curve of this nova (Fig. 1). Netzer (1975) called attention to the effect of self-absorption of Lyman and Balmer photons in high density nebulosity. Netzer (1975) showed that the ratio H/H from high density nebulosity are found in the range from 1.28 to 19.89 owing to the effect of self-absorption.
When the optical depths of Lyman ( ) and H( ) are very high, e.g. and , we have low ratios of H/H similar to those found on December 5 and 6. The intensity ratios higher than 4.0 are found in some conditions of higher optical depth of Lyman and/or lower optical depth of H (Netzer 1975). It may be very likely that the optical depths of Lyman and H were very high in the earliest stage, and the effect of self-absorption was important. The large variations of the ratio H/H could be explained with this effect. A similar variation of the ratio H/H was observed also in Nova (V1500) Cyg 1975 (Ferland 1978).
Radial velocities of selected absorption lines are given in Table 3, where Fe II is the mean radial velocity of Fe II lines and He I+N II is that of He I and N II lines. The observational errors in the individual measurements are about 100 km s-1. The radial velocities are given in the frame of the nova system of which the receding velocity is estimated as +46.3 0.5 km s-1 measuring the line center of the emission lines of [O III] and [N II] in the nebular stage. An absorption line at 6125 Å on December 5 is identified as the O I multiplet No. 10 (Moore 1959). The weighted mean of the wavelengths of the three contiguous lines of the multiplet is 6157.03 Å.
Figure 9 shows the mean radial velocities of the absorption components of H I, Fe II, and He I+N II. The radial velocities of H I, He I and N II rapidly increased (in negative sense) during the early decline stage. On the other hand, those of Fe II remained nearly constant or slightly decreased.
|Figure 9: Radial velocities of the absorption components of H I, He I, N II and Fe II lines.|
|Open with DEXTER|
Increasing outflow velocity has been observed in many previous novae (e.g. McLaughlin 1960). This phenomenon may represent the increasing of escape velocity due to the shrink of the photosphere (see, e.g. the model of Kato & Hachisu 1994). The absorption lines of Fe II might have been due to a remnant of the spectrum at light maximum.
Kiss & Thomson (2000) estimated the expansion velocity of the ejecta as 2000 km s-1 from their observations on December 8, but the outflow velocity largely increased after their observations. The mean radial velocity of He I and N II on December 17 was -2510 km s-1, and it seems that the velocity was still increasing (in negative sense) at that time. The outflow velocity may have increased up to about 2900 km s-1, because as shown later high velocity jets having radial velocities of about 2900 km s-1 were observed in the transition stage (Sect. 5.2).
There were two unidentified absorption lines in the spectra of the early decline stage. The wavelengths of the lines are given in Table 3. Since the wavelengths did not change so much, these lines may be related to the group of Fe II. If the radial velocities of these lines were the same as the mean of Fe II lines, their laboratory wavelengths should be 1.0 and 1.0 Å, respectively. We are, however, not able to find good candidates for these lines in the table of Moore (1959).
One possible interpretation for these lines may be that the former line is identified as He I 4471 blue-shifted by -1130 km s-1 on December 6 and -1150 km s-1 on December 17, and the latter is Na I D1+D2 5893 blue-shifted by -1260, -1270, and -1140 km s-1 on the respective days. It is, however, required a detailed work for the structure of the atmosphere of the nova to explain the reason why their radial velocities were so different from those of the main absorption components of the same lines.
We were not able to observe the nova in January 2000 because of its low celestial location. When the first spectrum in the new observational season was taken on February 6, 2000, this nova was in the transition stage. A large oscillation of luminosity, which is a typical phenomenon of the transition stage, was observed in the period from February to the middle of April 2000 (JD 2451570-1650) as seen in Fig. 1.
The spectroscopic observations in the transition stage were made on February 6, 21, March 7, 16 and 17. Figure 10 shows a tracing of the spectrum on February 6. The prominent emission lines are of H I, He I, He II, N II, N III, and [O III], while the emission lines of Fe II had disappeared. A weak emission line at 4245 Å is identified as [Fe II] 21F 4244.8 Å. This nova may be one of rare fast novae which show [Fe II] lines like CP Pup (Weaver 1944). The other lines of [Fe II] are blended with stronger lines of other elements.
|Figure 10: A spectrum of V1494 Aql on Feb. 6, 2000. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 11: A high dispersion spectrum of V1494 Aql on Feb. 21, 2000. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1. The dotted line shows a tracing of the spectrum multiplied by ten.|
|Open with DEXTER|
Table 4: Intensities of selected emission lines corrected by E(B-V)= 0.6 in the transition and nebular stages measured with the echelle spectrograph.
Table 5: Intensities of selected emission lines corrected by E(B-V)= 0.6 in the transition and nebular stages measured with the B&C spectrograph.
The first high dispersion spectrum with the Echelle spectrograph was taken on February 21, 2000, whose tracing is shown in Fig. 11. A part of the spectrum multiplied by ten is over-plotted as a dotted line to reveal the faint spectral features. There are prominent emission lines of H I, He I, He II, N II, N III, Si II, [O I], [O III], [N II], [Ca V] and [Fe VI]. The emission line at 6086 Å is usually identified as a blend of [Ca V] 6085.9 and [Fe VII] 6085.5. In this spectrum, however, the emission line may be due to [Ca V] 6085.9 alone, because no trace of [Fe VII] 5721 is seen. The emission lines of [Fe VII] were detected in September 2000 (Sect. 6.6). Intensities of prominent emission lines measured in the high dispersion spectra in the transition and the nebular stages are summarized in Table 4, and those measured in the medium dispersion spectra are given in Table 5. The effect of the interstellar extinction is corrected assuming E(B-V)= 0.6. Detailed process of the decomposition of blended lines is given in Sect. 6.1. The observational errors of the intensities of the prominent emission lines are about 10%, and those of the weak lines, which means that the intensity is of the order of , may be more than 50%. The intensities affected by larger errors are denoted by a colon.
During the transition stage four light maxima of the oscillation were observed at JD 2451587, 1606, 1621 and 1637 with an uncertainty of about 1 day. The interval between the maxima was 16.5 1 days and the mean amplitude of the oscillation was roughly 1.2 mag in V band. The observations on February 6, 21 and March 7 were made near light minima of the oscillation, while those on March 16 and 17 were made near the third maximum at JD 2451621.
Figures 12 and 13 show tracings of the high dispersion spectra taken on March 16 and 17, respectively. The spectral features near the light maximum were significantly different from those at light minima (Figs. 10 and 11). The emission complex at 4640 Å was already much stronger on March 16 (Fig. 12) than on February 21 (Fig. 11), and its further large growth (nitrogen flush) was observed on March 17 (Fig. 13). On the other hand, the emission line of He II 4686 weakened near the maximum of the oscillation and was embedded in the complex. The emission complex at 4640 Å showed two broad peaks on March 17 (Fig. 13), of which wavelengths were 4610 and 4647 Å. The former may be due to N II 4607.2 and 4613.9, while the latter may be a blend of N III 4634.2, 4640.6, 4641.9, C III 4647.4, 4650.2, 4651.4 and probably O II 4649.1 Å. In the red parts of the spectra the emission lines of N II appeared or strengthened at light maximum, hence He II and [Ca V] disappeared (Figs. 12 and 13). Similar spectral variations were known on the other oscillating novae GK Per and V603 Aql (see, e.g. McLaughlin 1943).
|Figure 12: A high dispersion spectrum of V1494 Aql on March 16, 2000. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 13: A high dispersion spectrum of V1494 Aql on March 17, 2000. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
The intensities of the forbidden lines also varied according to the oscillation. McLaughlin (1943) reported that the intensity of [N II] 5755 increased and that of [O III] 4363 decreased at light maxima of the oscillation of V603 Aql. Similar variations were seen on the present nova (Table 4). The intensity ratios of [O III] I(5007+4959)/I(4363) and [N II] I(6584+6548)/I(5755) suggest that the emitting region of [N II] lines had higher temperature or higher density than that of [O III] lines at the maximum of the oscillation. (Detailed arguments of these intensity ratios are given in Sect. 6.2.)
The profiles of N II 6167.8 and [N II] 5755 on March 16 are shown in Fig. 14. The emission line of [N II] 5755 shows well separated blue and red peaks and some narrow smaller peaks around line center. Many prominent emission lines show similar profiles (Figs. 12 and 13). On the other hand, the emission line of N II 6167.8 shows only a featureless broad single peak. Also the emission lines of N III may have had similar hazy profiles near the light maximum, because no discrete peak was seen on the emission complex at 4640 Å (Figs. 12 and 13). The emission lines of N II and N III seem to have been formed in a more compact and nearly spherically symmetric region.
|Figure 14: Profiles of [N II] 5755 and N II 6168 on March 16, 2000. The ordinate is an arbitrary logarithmic scale.|
|Open with DEXTER|
The emission component of H of V603 Aql at minimum of the oscillation in the transition stage showed two well separated peaks, whereas a broad single peak, which is rather similar to that of our N II 6167.8, was seen at maximum (McLaughlin 1956). Such a large variation of the profiles of H I lines was not observed in the present nova (Figs. 11-13).
High velocity broad emission components appeared in the blue and red sides of H and H on March 16 and 17 (Figs. 12 and 13), namely near the light maximum of the oscillation. Figure 15 shows tracings of the H region of the spectra taken on March 16 and 17. The radial velocities of the broad emission components in the frame of the nova system are given in Table 6. The observational errors are about 100 km s-1. The radial velocity of the red component of H was not measured, because it was blended with the emission line of [N II] 6584. The sum of the flux of the broad emission components of H was roughly 3% of that of the main emission component on March 16 and was about 6.5% on March 17.
|Figure 15: High velocity emission wings of H on March 16 and 17.|
|Open with DEXTER|
Table 6: High velocity emission components of H I lines.
It seems that high velocity jets were ejected at the light maximum of the oscillation. At far as we know, however, no previous nova showed a similar phenomenon during the transition stage. Even in the detailed description of the spectral variations of oscillating novae in the transition stage (McLaughlin 1943), such a high velocity emission component was never mentioned.
Rather similar high velocity emission components of H I and He II lines have been observed in super-soft X-ray sources, e.g. -6951 (Crampton et al. 1996) and (Becker et al. 1998). The radial velocities of the high velocity emission components of the former object were 3600 km s-1 (Crampton et al. 1996) and those of the latter were about 800-900 km s-1 (Becker et al. 1998). The high velocity emission components of these objects, however, lasted at least several months keeping nearly the same velocities. On the other hand, the intensity and radial velocity of the high velocity emission components of the present nova were rapidly variable.
At the present time, our knowledge of the high velocity emission components of classical novae seems to be too poor to understand their properties. Detailed observations of the spectral variations in transition stage of novae in the future are required.
The oscillation of luminosity in the transition stage terminated at the middle of April 2000 (Fig. 1). A tracing of the high dispersion spectrum taken on April 23 is shown in Fig. 16. Notable differences with respect to the spectrum on February 21 (Fig. 11) are growth of [O III] and [Ca V] lines relative to H I.
Almost all prominent emission lines showed saddle-shaped profiles in the transition (Figs. 10-13) and nebular (Fig. 16) stages. Therefore, it is not possible to decompose blended lines with the usual Gaussian fitting. In this paper, we estimated the intensities of individual components of blended lines with the following method.
Since many lines in a spectrum showed nearly the same profile, we used the profile of H as the standard, and measured the intensities in the blue and red quarters of the emission component. For example, the wavelengths of the two peaks of H on the spectrum shown in Fig. 16 are 4843.5 and 4880.0 Å, respectively. The full width at zero intensity of the emission component is about 1.4 times larger than the separation of the two peaks, that is the wavelengths at zero intensity are 4836.2 and 4887.3 Å. We measured the intensities in the blue quarter 4836.2-4849.0 and in the red quarter 4874.5-4887.3 of the emission component, then compared them with the total intensity. The ratio of the intensity in the blue quarter to the total one is 0.23 and that in the red quarter is 0.24 in the present case. Fortunately, the blue or red quarter of many blended lines were free from blending. In such cases, we were able to estimate the intensities of individual components of blended lines by this method. This method was not applied to N II and N III lines in the transition stage, because their profiles were different to that of H (Figs. 12-14). The intensities of the prominent emission lines in the transition and nebular stages are given in Tables 4 and 5.
|Figure 16: A high dispersion spectrum of V1494 Aql on April 23, 2000. The unit of the ordinate is 10-10 erg cm-2 s-1 Å-1. The dotted line shows a tracing of the spectrum multiplied by ten.|
|Open with DEXTER|
As mentioned in Sect. 4.2, the high intensity ratio of H/H lasted during the transition stage, then gradually decreased (Fig. 8). The ratio H/H observed on April 26 and later (Tables 4 and 5) is rather close to those expected in the recombination theory (Hummer & Storey 1987). Probably, the effect of the self-absorption (Netzer 1975) became weak in that period. The properties of the ejecta, therefore, are analysed using the spectra taken on and later than April 26, 2000.
Figure 17 shows the profiles of the emission lines of H,
[N II] 5755 and [O III] 5007 on May 15. As seen in the figure, these three
lines have the same detailed structures. It seems that these
lines were formed in nearby regions. The uniform electron temperature and
density of the emitting regions of these lines, therefore, could be a good
approximation for the ejecta of this nova in the nebular stage. If this is
the case, it is easy to obtain the electron temperature and density from
the intensity ratios of the nebular lines to the auroral lines of [O III]
and [N II]. Seaton (1975) gave the following formulae.
|Figure 17: Profiles of H, [N II] 5755 and [O III] 5007 on May 15, 2000. The ordinate is an arbitrary logarithmic scale.|
|Open with DEXTER|
Table 7: Physical properties of ejecta.
The electron temperature was nearly constant at 10 700 K during the nebular stage. The variation of the electron density with time is shown in Fig. 18. As seen in the figure, the electron density decreased smoothly with a decline rate of , where t is number of days from light maximum. This decline rate is much lower than that expected in the spherically symmetric expansion . The ejecta might have a form of a ring and, probably, polar blobs such as those of V603 Aql (Weaver 1974), HR Del (Hutchings 1972), DQ Her (Martin 1989) etc.
|Figure 18: Electron density of the ejecta in the nebular stage.|
|Open with DEXTER|
Another possible cause of the low decline rate is a continuous mass outflow in the nebular stage. The model of Kato & Hachisu (1994) shows that if the mass of the white dwarf is less than 1.15 , significant mass outflow will continue even more than 150 days after the explosion. Probably also this effect could contribute to the low decline rate.
As mentioned in Sect. 5.1, the intensity ratios of [O III] and [N II] lines at the light maximum of the oscillation in the transition stage were rather peculiar. If we assume that the electron temperature was 10 700 K in the transition stage, the intensity ratios of the lines on March 16 and 17 give higher densities in the emission region of [N II] than in the [O III] region. On the other hand, if we extrapolate the decline rate of the electron density in the nebular stage to the transition stage, the same intensity ratios give higher temperature in the [N II] region than in the [O III] region. In any case, the condition of the ejecta at the light maximum of the oscillation may have been different from the thermal equilibrium.
The helium abundance of the ejecta is estimated from the intensity ratios of He I 5876/H and He II 4686/H.
In the previous sub-section, we have found
106 cm-3 and
K for the ejecta in the nebular
stage. The effect of collisional excitation of He I 5876 can not be
neglected in such a nebulosity (Clegg 1987; Peimbert &
Torres-Peimbert 1987). Here, we use the formula of Peimbert &
Torres-Peimbert (1987) to estimate the ratio of
collision/recombination of He I 5876. The results are given as He I 5876 C/R
in Table 7. The effective recombination coefficients of H,
He I 5876
and He II 4686 at
106 cm-3 are:
Following Osterbrock (1989), we have a relation between the
intensity of the H emission line and the mass of hydrogen in the
emitting nebula of pure hydrogen as:
If we assume that the ejecta has a uniform density, the mass of hydrogen
in the ejecta will be given as:
The radius of the ejecta on April 26 may be about 3.6 1010 km, where we assumed the expansion velocity of the ejecta to be 2900 km s-1. Using the mass and the electron density given in Table 7, we have for April 26. The error in is mainly due to the uncertainty in the expanding velocity. If the error in the expanding velocity is 200 km s-1, that in is about 0.01. This factor was 0.022 on May 15 and 0.016 on June 9. These small factors also suggest that the ejecta was not spherically symmetric, but may have had a ring-like shape.
The low decline rate of the electron density (Sect. 6.2) and the small factor obtained in the previous sub-section suggest that the ejecta may have had a ring-like form in the nebular stage. In addition to these results, the widths of the emission lines seem to support the model of the ring-like form of the ejecta.
The outflow velocities measured with the blue-shifts of the absorption lines in the early decline stage were larger than 2500 km s-1 (Fig. 9) and probably reached up to 2900 km s-1 (Sect. 4.3). If the ejecta were expanding with such velocities, the full widths of the emission lines in the nebular stage should have been about 5000-6000 km s-1. On the other hand, the FWHMs of the emission component of H in the nebular stage were about 2700 km s-1 (Table 4) and the full width at zero intensity, which was measured on April 23 (Sect. 6.1), was 3200 km s-1. The widths of the other prominent emission lines were nearly the same.
The disagreement between the radial velocities of the absorption lines and the widths of the emission lines could be explained if we assume that the ejecta in the nebular stage formed a ring which was expanding with the velocity of 2900 km s-1, but was inclined by about 30 degrees from the line of sight.
At the present time, three years after the explosion, the radius of the ejecta may be of order of 0.1 arcsec. This radius is large enough to be measured from ground based instruments or by the Hubble Space Telescope. Further works are awaited.
A tracing of a red spectrum taken on September 9 is shown in Fig. 19, where the emission lines of [Fe VII] and the coronal line [Fe X] 6347 are seen. Almost all prominent emission lines in the optical region (this paper) and the coronal lines in the infrared region (Rudy et al. 2001) showed double-peaked profiles, but the profile of [Fe X] 6374 looks different. Its strange profile may be due to a blending with Si II 6347. The FWHM of H was about 2700 km s-1 throughout the nebular stage (Table 4), and the widths of the other prominent emission lines were nearly the same. If the widths of Si II 6347 and [Fe X] 6375 were also about 2700 km s-1, the red part of Si II and the blue part of [Fe X] should overlap each other producing a single strong peak between them. A similar profile of blended lines is seen in the case of [Fe VI] 5147 and 5177 on April 23 (Fig. 16)
Figure 20 shows a tracing of our last spectrum taken on September 16, 2000, where [O III] 5007 is much stronger than H. Like the spectrum taken on September 9 (Fig. 19), some emission lines of [Fe VII] and the coronal line [Fe X] 6347 are seen. No other coronal line in the optical region, e.g. [Fe XIV] 5304, is seen in this spectrum. The same condition is known also in Nova (GQ) Mus 1983 (Krauter & Williams 1989). On the other hand, many coronal lines in the infrared region, e.g. [Si VII], [Si X], [Ca VIII], [Al IX], [S IX] etc., were observed in July 2000 (Venturini et al. 2000) and July 2001 (Rudy et al. 2001).
|Figure 19: A spectrum of V1494 Aql on September 9, 2000. The unit of the ordinate is 10-12 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
|Figure 20: A high dispersion spectrum of V1494 Aql on September 16, 2000. The unit of the ordinate is 10-12 erg cm-2 s-1 Å-1.|
|Open with DEXTER|
The unidentified emission line at 6200, which was detected on CP Pup (Sanford 1945), is seen on this nova (Figs. 19 and 20). Since this line appeared in a later nebular stage, it may relate to a highly ionized ion. Its precise laboratory wavelength should be 6199.5 0.5 Å.
In this paper, a new distance 1.6 0.2 kpc is derived for the nova V1494 Aql. The mass of the ejecta is estimated as 6.2 1.4 and the helium abundance as N(He)/N(H) =0.13 0.01. The electron temperature lasted nearly constant at 10 700 K during the nebular stage, at least until the end of our monitoring in September 2000. The electron density decreased as , where t is number of days from light maximum. This low decline rate suggests that the ejecta had a ring like form and that probably a significant mass outflow continued in the nebular stage.
This nova seems to be similar to V603 Aql in many aspects, e.g. the oscillation of luminosity in the transition stage, the spectral variations and the form of the ejecta. However, the high velocity jets observed around the light maximum of the oscillation in the transition stage have not been found in any previous nova. Careful spectroscopic observations in the transition stages of other novae are required to confirm the existence and better define the properties of the high velocity jets.
We are grateful to Profs. I. Hachisu and M. Kato for the useful discussions and comments, and to Prof. R. Barbon for the careful reading of the manuscript and useful suggestions. Thanks are also due to Dr. T. Kato, Dr. D. Nogami, Mr. S. Kiyota, and many other professional and amateur astronomers who supplied photometric data of V1494 Aql on the VSNET. HHE gratefully thanks both Italian Government Ministry of Foreign Affairs and the Research Fund of the University of Istanbul for financial support under a scholarship and the project 1506/280700.
Table 1: Log of spectroscopic observations of V1494 Aql.