The Complex Structure of Classical Nova Shells

J. J. Johnson, T. E. Harrison, and D. M. Leeber

New Mexico State University

Abstract

We examine the changing structure of the shell around Nova Cassiopeiae 1995. Its continuing evolution is marked by changes in line profiles, indicative of changing structure. The different atomic species show differing profile variations, thereby indicating that they arise from different locations in the circumnova nebula. We thus have further evidence that the structure of this nova's shell is more complex than a spherical shell or even a single toroid around the nova.

Introduction

Nova Cas 1995 (V723 Cas) was discovered on August 24, 1995 by K. Hirosawa (IAUC 6213). Its subsequent behavior showed it to be a slow nova, similar to HR Del. To date, it has not returned to its pre-maximum magnitude of 18-19, and has been hovering at V=13.5 since June 1999.

The early evolution of the nova was unusual for the length of time between discovery and visual maximum (115 days). The spectra before maximum showed a stellar-like continuum with Balmer and Paschen lines varying between emission and absorption. After maximum, the spectra showed numerous emission lines of low-ionization species on a flat continuum. It wasn't until July 1997 that Nova Cas 1995 finally entered the nebular stage (IAUC 6703). We have extensive spectra and spectropolarimetry of the early stages (Johnson, et al. 2000) which showed that early in the evolution of this slow nova, there was significant structure in the ejecta. This poster discusses observations of Nova Cas after entering the nebular stage.

The Observations

We obtained observations of Nova Cas 1995 at Apache Point Observatory (APO) using the Double-Imaging Spectrograph (DIS) in high-resolution mode. In this mode, the blue resolution is 1.6/pixel and the red resolution is 1.3 /pixel. The spectra were centered at 4650 and 6600, respectively, in order to include H, H, and H, as well as lines of He I, He II, and [O III]. We rotated the position angle of the slit in order to investigate any differences there might be in the line profiles, as Solf (1983) did for HR Del. The details of the observations are shown in Table 1. We made several exposures, varied exposure times, and the position of the spectra on the CCD chip in order to ensure that the variations in line profiles are not due to chip defects or S/N effects.

Table 1
Date	Position Angles
Oct. 11, 1997	0, 45, 90, 135
Dec.15, 1997	0, 90
Nov. 5, 1998	0
Nov. 6, 1998	0, 45, 90, 135

The Line Profiles

We are interested in three possible effects: the variations of the line profiles from species to species, variations with slit position angle, and changes of line profiles with time. As a side consideration, we will also investigate the relative strengths of lines as the nova shell evolves. The main lines we see are H, H, H, He I 4471 and 6678, He II 4686, [O III] 4363 and 5007, [O I] 6300 and 6364, O I 7002, and [Ar V] 6435 and 7006. Some of these lines are blended (such as O I and [Ar V]), and will be treated cautiously. The lines are shown in Figures 3-8. The left panels are our measurements at 0, right panel at 90, and date increases from top to bottom. Figure 3 is H-alpha. Figure 4 is H-beta. Figure 5 is H-gamma plus [O III] 4363. Figure 6 is [O III] 5007. Figure 7 is He I 4471 and He II 4542. Figure 8 is He II 4686.

Several results are immediately obvious.

1) Virtually all lines are double-peaked.

2) Though sometimes the peaks are of equal strength, in most lines one peak is stronger.

3) The dominant peak changes in different lines, at different position angles, and with time.

We now investigate these changes in more detail.

The Analysis

To begin the analysis, we broke the lines into two separate components, using the data analysis routines in IRAF. Most lines were reproduced well by the sum of two Gaussians (Figure 2).

In all these cases, the lines were separated by 350 km/s.

The values ranged from 320 km/s to 380 km/s but were remarkably consistent from observation to observation, and line to line. (The difference is less than our resolution).

The FWHMs of the separate components ranged from 4.5 to 5.7 generally in the blue, and 6.5 to 7.9, i.e. 300 km/s. Again, the values were consistent from observation to observation.

The velocity of expansion did not vary, nor did the velocity width of the shell change. What changed was the relative strengths of the two components.

The equivalent widths of the two components of the lines showed the changes with time.

1) In all 3 Balmer lines, the red peaks were much stronger then the blue peaks in 1997, but in 1998 the blue peak strengthened to become nearly equal or slightly stronger (in the case of H), at all position angles.

2) [O III] 5007 also evolved from having the red peak stronger than the blue peak, to having roughly equal peaks. However, [O III] 4363 always had the blue peak stronger.

3) He II 4686 always had the blue peak stronger, though Figure 7 shows that He II 4542 showed a more complicated behavior (given the low S/N of the line in 1997, those measurements are regarded with some suspicion).

4) He I 6678 showed equal peaks at all times and all position angles. He I 4471 showed a tendency for the blue peak to dominate, but that was position angle dependent.

Discussion

It is relatively easy to explain profiles with peaks of equal strength via an expanding uniform optically thin shell, or disk. Having a stronger blue peak in a double-peaked line is also explained straightforwardly, by assuming that the red peak is being partially obscured by intervening circumnova material.

However, having the red peak stronger requires a bit more complicated geometry. There must be more material and/or more favorable excitation conditions in the area of the nebula generating that portion of the emission line.

Analysis also shows that the strengths of some lines (e.g. He II 4542 in 11/98) vary with position angle, again suggesting that the geometry is more complex than an expanding shell with obscuring material.

Both Solf (1983) and Hutchings (1972, see Figure 9) saw similar effects in HR Del, the classical nova most similar to Nova Cas 1995. Though they had much higher spectral resolution and a closer, brighter nova to work with, we can apply their models to our spectra. Figure 10 is the operative figure from Solf (1983) showing the different components he required to replicate the complex line profiles of HR Del. Hutchings also used a truncated cone and equatorial ring combination. In the case of HR Del, he even invoked a triple polar cone in order to reproduce the complex line profile.

It is clear that the complex structure of the line profiles observed from Nova Cas 1995 is either a combination of two or three separate emission regions, or the ejecta must be extremely clumpy. The latter would explain why the red peak is sometimes stronger if the ejecta moving away from us was denser than that moving toward us. We are experimenting with various combinations of polar cones and disks at different viewing angles to investigate the models of Hutchings and Solf.

Conclusions

The structure of the shell around Nova Cas 1995 is more complex than an expanding spherical bubble. There are definite differences in the density of the ejecta both along the line-of-sight and at different position angles. The spectro-polarimetry obtained early in the evolution showed that the structure then was spherical and likely to be bipolar. These new data show that the circumnova material is still not spherically symmetric, nor is it homogeneous. Models of abundances in nova shells should take into account the variations of density and excitation conditions we see in Nova Cas 1995.

This work is supported by NSF award number AST-9806114.