IC 4776 and its binary central star

The planetary nebula IC 4776 and its post-common-envelope binary central star

Paulina Sowicka, David Jones, Romano L. M. Corradi, Roger Wesson, Jorge García-Rojas, Miguel Santander-García, Henri M. J. Boffin, and Pablo Rodríguez-Gil
Nicolaus Copernicus Astronomical Center, Bartycka 18, PL-00-716 Warsaw, Poland
Instituto de Astrofísica de Canarias, E-38205 La Laguna, Tenerife, Spain
Departamento de Astrofísica, Universidad de La Laguna, E-38206 La Laguna, Tenerife, Spain
Department of Physics and Astronomy, University College London, Gower Street, London WC1E 6BT
Observatorio Astronómico Nacional (OAN-IGN), C/ Alfonso XII, 3, 28014, Madrid, Spain
GRANTECAN, Cuesta de San José s/n, E-38712 , Breña Baja, La Palma, Spain
European Southern Observatory, Karl Schwarzschild Strasse 2, 85748 Garching, Germany
E-mail: paula@camk.edu.pl
Accepted XXX. Received YYY; in original form ZZZ
Abstract

We present a detailed analysis of IC 4776, a planetary nebula displaying a morphology believed to be typical of central star binarity. The nebula is shown to comprise a compact hourglass-shaped central region and a pair of precessing jet-like structures. Time-resolved spectroscopy of its central star reveals periodic radial velocity variability consistent with a binary system. While the data are insufficient to accurately determine the parameters of the binary, the most likely solutions indicate that the secondary is probably a low-mass main sequence star. An empirical analysis of the chemical abundances in IC 4776 indicates that the common-envelope phase may have cut short the AGB evolution of the progenitor. Abundances calculated from recombination lines are found to be discrepant by a factor of approximately two relative to those calculated using collisionally excited lines, suggesting a possible correlation between low abundance discrepancy factors and intermediate-period post-common-envelope central stars and/or Wolf-Rayet central stars. The detection of a radial velocity variability associated with binarity in the central star of IC 4776 may be indicative of a significant population of (intermediate-period) post-common-envelope binary central stars which would be undetected by classic photometric monitoring techniques.

keywords:
planetary nebulae: individual (IC 4776, PN G002.013.4) – binaries: spectroscopic – stars: mass loss – ISM: jets and outflows
pubyear: 2017pagerange: The planetary nebula IC 4776 and its post-common-envelope binary central starA

1 Introduction

Planetary nebulae (PNe) are the intricate, glowing shells of gas ejected by low- and intermediate- mass stars at the end of their asymptotic giant branch (AGB) evolution which are then ionized by the emerging pre-white dwarf core. With some 80 per cent of all PNe showing deviation from spherical symmetry (Parker et al., 2006), it has proven impossible to understand their structures in terms of single star evolution (Soker, 2006; Nordhaus et al., 2007; García-Segura et al., 2014), with binarity frequently invoked to explain their diverse, often strongly axisymmetrical morphologies (De Marco, 2009; Jones & Boffin, 2017).

While a lower limit to the close-binary central star fraction is well constrained (at 20%) by photometric monitoring surveys (Miszalski et al., 2009a), it is insufficient to explain all aspherical PNe. The remaining aspherical PNe are generally understood to be the products of mergers, wider binaries and/or weaker binary interactions (i.e. the engulfment of a Jovian mass planet; De Marco & Soker, 2011). This hypothesis is supported by common-envelope (CE) population synthesis models, which predict a significant number of post-CE binaries with orbital periods of several days to a few weeks (e.g. Han et al., 1995) while almost all of the known post-CE central stars have periods less than one day (Jones & Boffin, 2017).

The lack of known post-CE central stars with intermediate periods (only three are known to have periods greater than three days, Méndez & Niemela, 1981; Manick et al., 2015; Miszalski et al., 2017) is, perhaps, not unreasonable given that the majority of the effort has been focused on photometric monitoring which becomes particularly insensitive at these longer periods (De Marco et al., 2008; Jones & Boffin, 2017). Recent work, however, has shown that targeted radial velocity monitoring can begin to reveal these missing binary systems (Miszalski et al., 2017; Jones et al., 2017). Constraining this population is of particular interest given that observations of “naked” (i.e. those with no surrounding PN) white dwarf plus main sequence binaries find a similar dearth of intermediate period systems (Nebot Gómez-Morán et al., 2011), strongly indicating that the lack of known systems is not purely an observational bias. Understanding to what extent this population is truly absent (rather than just difficult to detect) will greatly further our understanding of the common envelope process itself (Toonen & Nelemans, 2013).

Figure 1: FORS2 narrowband imagery of IC 4776. Both H images include the light of [N ii], and the central region of the rightmost H image is heavily saturated but most clearly shows the structure of the faint, precessing jets. All images measure 0.5′0.5′, north is up and east is left.

IC 4776 is a relatively bright, small planetary nebula comprising a central, bipolar, hourglass-like structure and an extended, jet-like structure revealed for the first time by the images presented in this paper (see Figure 1 and Section 2.1). Both jets and hourglass structures have been strongly linked to binary central stars (Miszalski et al., 2009b). The nebula has been shown to display dual-dust chemistry (i.e. features associated with both carbon- and oxygen-rich dust; Perea-Calderón et al., 2009; Górny et al., 2010), which has also been linked to a possible binary evolution of the nebular progenitor (Guzman-Ramirez et al., 2014). The spectral type of its central star is unclear but has been classified by various authors as [WC] (e.g.  Aller & Keyes, 1985). The apparent presence of narrow emission lines has also earned it a wels classification (Tylenda et al., 1993), now generally not considered a valid classification given that in many cases the emission lines themselves do not originate from the stars but from their host nebulae (Basurah et al., 2016) or are the product of the irradiation of a main-sequence companion (Miszalski et al., 2011b).

Based on the likelihood that it hosts a binary central star, IC 4776 was selected for further study as part of a programme to search for binary central stars through time-resolved radial velocity monitoring. Here, we present the results of that radial velocity study, revealing variability consistent with an intermediate period, post-CE binary. We furthermore present spatio-kinematic and chemical analyses of the nebula itself, in order to further constrain the relationship between the nebula and the probable central binary.

2 Nebular morphology and kinematics

2.1 FORS2 narrowband imagery

Narrowband imagery of IC 4776 was obtained using the FORS2 instrument mounted on ESO’s VLT-UT1 (Appenzeller et al., 1998). Exposures were acquired in the following emission lines: H+[N ii] (5s exposure time on 2012 September 5, 30s on 2016 June 11), [O ii3727 (20s 2012 September 3), and [O iii5007 (5s 2012 September 5). In each case the seeing was better than 1″. The debiased and flat-fielded images are presented in Figure 1.

In all three bands, the nebula shows a similar morphology, namely an “X"-shape, which is most prominently visible in the light of [O ii3727. At all three wavelengths the central shell appears the brightest, while only at [O ii3727 it is possible to discern its detailed structure. This hourglass morphology is extremely similar to that of MyCn 18, a PN often hypothesised to have originated from a binary interaction owing to its similarity to some nebulae surrounding symbiotic binary stars (Corradi & Schwarz, 1993). MyCn 18, in addition to its central hourglass, displays a system of high velocity knots which may well be analogous to the jets observed in IC 4776 (Bryce et al., 1997; O’Connor et al., 2000), further highlighting the similarities between the two nebulae. The jets of IC 4776 are evident in all filters but most prominently in the light of H where their remarkable structure is clearest. The evident curvature of the jets is a strong indication of precession, which can be a natural product of a central binary system (Raga et al., 2009). Such precessing jets have been observed in several close-binary PNe, e.g. Fg 1 (Boffin et al., 2012), ETHOS 1 (Miszalski et al., 2011a), and NGC 6337 (García-Díaz et al., 2009; Hillwig et al., 2016), providing a solid observational link between jets and binarity.

2.2 FLAMES-GIRAFFE integral field spectroscopy

Spatially-resolved, high-resolution, integral-field spectroscopy of IC 4776 was obtained on August 8 2013 using FLAMES-GIRAFFE mounted on ESO’s VLT-UT2 (Pasquini et al., 2002). The seeing during the observations did not exceed 1″. The ARGUS integral field unit, employed in its standard ‘1:1’ scale mode (resulting in a total aperture of 11.5 x 7.3″ sampled by an array of 2214 0.52″0.52″ microlenses), was used to feed the GIRAFFE spectrograph setup with the H665.0/HR15 grating (providing a resolution of R30 000 in the range 6470Å 6790Å). ARGUS was oriented at P.A. = 38 (to coincide with the symmetry axis of the nebula) and with only two telescope pointings, offset by about 10″ to each other, we were able to obtain a complete coverage of the nebula (which is 20″ long including the faint extended emission), as shown in Figure 2. Two exposures of 245-s were obtained for each of the two pointings, with the resulting data reduced and combined to produce a single, wavelength-calibrated data-cube for each pointing using the standard ESO-GIRAFFE pipeline. Finally, the data from the two pointings were corrected to heliocentric velocity and combined to form a single data-cube using specially written python routines.

In the wavelength range of the ARGUS data there are several spectral lines: H, [N ii6548,6584 and the [S ii6717,6731 doublet. In order to probe the kinematics and morphology of the nebula, we focus on the [N ii6584 line given that it is brighter than [S ii6717,6731 and has a lower thermal width than H.

Figure 2: The ARGUS-IFU coverage of IC 4776. Each image measures 1′ 1′. North is up, East to the left. Left: the two pointings merged together, overplotted on FORS2 image in the light of [O II]. Right: the two pointings shown separately, overplotted on FORS2 image in the light of H in order to highlight the spatial coverage of the jets.
Figure 3: Velocity channel map for the [N ii6583.45 emission from IC 4776. Velocities are corrected to Heliocentric, and each channel represents the summed emission of three “slices” of the reduced data cube, corresponding to a velocity bin of 6.83 km s. The logarithmic display limits for each channel are calculated individually in order to maximise the features visible on each map.

Channel maps for the [N ii6584 line are shown in Figure 3. Each channel shows a cross-section through the nebula at a different velocity relative to the rest velocity of [N ii6584, starting from the top left-hand side channel being the most blue-shifted (velocity of about km s), ending on the bottom right-hand side channel being the most red-shifted (velocity of about 137 km s). Each channel is a sum of three slices of the reduced cube where the wavelength step between slices is approximately 0.05Å (i.e., each channel represents a range of 0.15Å6.83 km s at [N ii6584). The brightness/contrast is calculated individually for each channel using a logarithmic function in order to display the maximum amount of structural information possible.

The systemic velocity of the nebula has been previously found to be =16.30.6 km s (Durand et al., 1998), and around this velocity the channel maps present two circular structures joined at the position of the central star. This can be interpreted as the contribution of the two lobes of an inclined bipolar shell, as inferred from the imagery presented in Section 2. Further from this central velocity, one lobe begins to dominate, just as would be expected from an inclined hourglass. Yet further still from the systemic velocity (e.g. 20km s), a second component begins to contribute significantly to the observed emission in each channel map, initially appearing close to the central star and then moving away at velocities furthest from the systemic velocity. This emission feature clearly originates from the jets, presenting with a similarly curved appearance at the largest velocities. The jets themselves also present clearly greater velocities than the central shell (the maximum extent of which is at roughly 35km s), reaching more than 100km s away from the systemic velocity. Furthermore, there is an evident asymmetry in the jet velocities given that the blue-shifted jet reaches roughly 130km s, while the red-shifted component only reaches as far as 100km s. Figure 2 reinforces this idea as in the high contrast image the Northern jet extends beyond the region encompassed by the FLAMES-GIRAFFE observations (centred on the central star), while the Southern jet is apparently well covered. Given the clear precession visible in both the imagery and IFU spectroscopy it is not possible to derive a simple correlation between line-of-sight velocity and distance from the central star which would further constrain the asymmetry.

A 2 spaxel region surrounding the central star was collapsed (equivalent to a slit-width of 1.04″) in order to produce a simulated longslit spectrum. The position-velocity array of this simulated longslit is shown in Figure 4. The position-velocity array shows a clear “X”-shape, typical of such bipolar, hourglass structures, as well as end-caps corresponding to the brightest regions of jet emission (further emission from the jets can be seen with increased contrast which blows out the central nebula).

Figure 4: Simulated position velocity corresponding to an [N ii6583.45Å longslit, with width 1″, placed along the symmetry axis of the nebula.
Figure 5: Synthetic images of the shape model of IC 4776 at the derived inclination of 42° (left) and at an inclination of 90° (right). The central image is the FORS2 [O ii] image (reproduced from Fig. 1 for direct comparison). Each image measures 0.25′0.25′.

2.3 Shape modelling

A spatio-kinematic model of the central hourglass of IC 4776 was constructed based on the narrowband imagery and spatially-resolved spectroscopy presented above. The shape software was used, following the standard workflow presented in Steffen et al. (2010), varying both morphological and kinematic parameters over a wide-range of parameter space with the best fit being determined from a by-eye comparison of synthetic images and velocity channel maps to the observations. The jets themselves are too complex to be treated by a simple spatio-kinematic model, and their detailed modelling is reserved for a subsequent paper (Santander-García et al., in prep). Although, it is worth noting that, in principle, the data are of sufficient quality to derive the precession properties of the jets, which can then be related to the parameters of the central binary star (e.g. Raga et al., 2009; Boffin et al., 2012) at the time the jet was launched (constrained by the kinematical age of the jet).

The best-fitting shape model, as expected from appearance of the nebula in the narrowband imagery, comprises an hourglass-like structure whose symmetry axis is inclined at 42°4° with respect to the plane of the sky (see Figure 5). The velocity structure of the nebula is found to be well represented by a flow-law whereby all velocities are radial from the central star and proportional to the distance from the central star. Such flow-laws, often referred to as “Hubble flows”, are generally taken to imply an eruptive event in which the majority of the nebular shaping occurred during a brief time period. However, Steffen et al. (2009) showed that while deviations from a “Hubble flow” should be appreciable and easily observed in PNe formed via the classical interacting stellar winds model (Kwok et al., 1978; Kahn & West, 1985), such a simple flow law will provide a good first order estimation of the overall morphological and kinematic properties of the nebula. Particularly given the relatively low spatial resolution of the FLAMES-GIRAFFE observations c.f. the size of IC 4776, we are unable to constrain the presence (or not) of such deviations from a “Hubble flow”.

The systemic velocity of the nebula is found to agree well with the literature value of 16.30.6 km s (Durand et al., 1998), and was fixed in the modelling. The kinematical age of the nebula is found to be approximately 300 yr kpc (with a significant uncertainty due to the small size of the nebula). While there is significant variation in distance determinations for IC 4776 in the literature, the kinematical age would imply a rather young nebula at all but the very largest distances. The distance of 1.01 kpc determined by Phillips & Pottasch (1984) would seem to imply an impossibly young nebula, while the distances derived by Stanghellini et al. (2008, 4.97 kpc) and Frew et al. (2016, 4.441.27 kpc) would give a more reasonable age of 1500 years.

3 Nebular chemistry

High-resolution longslit spectroscopy of the nebula IC 4776 was acquired on August 9 2016 using the Ultra-violet and Visual Echelle Spectrograph (UVES) mounted on ESO’s VLT-UT2 (Dekker et al., 2000). Exposures were acquired using both arms of the spectrograph with the following set-ups. Using Dichroic 1 (3 150s exposures, 1 20s exposure), the blue-arm was set to a central wavelength of 3460Å (with the standard HER_5 blocking filter) while the red-arm was set to a central wavelength of 5800Å (with the standard SHP700 blocking filter). Using Dichroic 2 (3 300s exposures, 1 60s exposure, 1 20s exposure), the blue-arm was set to a central wavelength of 4370Å (again with the standard HER_5 blocking filter) while the red-arm was set to 8600Å (with the standard OG590 blocking filter). A slit-width of 2.4″ was employed, and the seeing during the observations was approximately 1.5″. The spectra were reduced using the standard UVES pipeline and flux-calibrated using 300s exposures of the standard star LTT 7987 taken with the same set-ups directly following the observations of IC 4776. Collectively, the spectra provide near-continuous wavelength coverage from 3000–10 000Å (see Figure 10). To construct the final spectrum, we took the median of the flux at each wavelength from the longest exposures in each setup, and where lines were saturated, replaced the affected ranges with values from the shorter exposures in which saturation was not an issue.

Emission line fluxes were measured using the alfa code (Wesson, 2016), which optimises parameter fits to the line profiles using a genetic algorithm following the subtraction of a global continuum. The effectiveness of alfa in measuring PN line fluxes has previously been demonstrated in several papers (e.g. Jones et al., 2016), and is further demonstrated in Figures 7 and 10. alfa assumes that all lines have a Gaussian profile. At the high resolution of the UVES spectra, velocity structure is evident, and lines of different ionisation have different profiles. For the purposes of chemical analysis, we binned the spectra by a factor of 10 in wavelength, resulting in Gaussian profiles for all lines.

353 emission lines were measured, of which 329 were resolved and 24 were blends of two or more lines. Échelle spectra may contain many spurious features due to bleeding of strong lines from adjacent orders. In general, these do not affect the analysis, as ALFA only attempts to fit features close to the wavelengths of known emission lines. Spurious features not fitted by ALFA may be seen in panels (a) and (b) of Figure 7 (though they are also present in all panels of Figure 10). However, in a few cases, the bleeding may be blended with nebular emission. The O ii recombination line at 4089Å is one such case, and this impacts the estimate of the temperature and density from the ratios of O ii lines, as discussed below. Table 5 lists the observed and dereddened fluxes measured for all emission lines measured in the spectra, together with their 1 uncertainties.

Physical conditions as well as chemical abundances were calculated from the emission line fluxes using the neat code (Wesson et al., 2012). The code uses Monte Carlo techniques to propagate uncertainties in line flux measurements into the derived quantities (for full details please see Wesson et al., 2012). The physical parameters determined for IC 4776 are listed in Table 1.

c(H =
([O ii]) (cm) =
([S ii]) (cm) =
([Cl iii]) (cm) =
([Ar iv]) (cm) =
(BJ) (cm) =
(PJ) (cm) =
([O ii]) (K) =
([S ii]) (K) =
([N ii]) (K) =
([O iii]) (K) =
([Ar iii]) (K) =
([S iii]) (K) =
(BJ) (K) =
(PJ) (K) =
(He 5876/4471) (K) =
(He 6678/4471) (K) =
(O ii ORLs) (K) =
(O ii ORLs) (cm) =
Table 1: Extinction, temperatures and densities of IC 4776 as derived using the neat code, as well as the temperatures and densities used for the determination of the chemical abundances listed in Tables 2 and 3.

The extinction towards the nebula was estimated from the flux-weighted average of the ratios of H, H and H to H, and the Galactic extinction law of Howarth (1983). H and H are in the same spectrum as H, while H is in a separate spectrum. The three lines give consistent estimates of the logarithmic extinction at H, c(H), which is 0.220.03.

The nebular density is estimated using several standard collisionally excited line (CEL) diagnostics, as well as from the Balmer and Paschen decrements, and the ratios of O ii recombination lines. The CEL diagnostics give relatively high densities of 10000–30000 cm; the Paschen and Balmer decrements also imply high densities, of 10–10 cm, although with larger uncertainties (see Figure 6). The O ii recombination line ratios imply a lower density of 3000 cm, with the caveat that the 4089 Å line flux on which the ratios rely may be overestimated due to bleeding from adjacent orders. If its flux is overestimated, the density derived is underestimated.

Figure 6: Observed intensities (relative to H=100) of high-order Balmer (a) and Paschen (b) lines as a function of the principal quantum number . The dashed curves show the predicted intensities of the lines as a function of electron density, at a temperature of K as derived from the [O iii] nebular to auroral line ratio. The plots highlight the general tendency towards larger densities (10–10cm), consistent with those measured from CEL diagnostics. Note that H14 is blended with a line of [S iii] and, as such, its flux is overestimated. P13 and P16 lie close to the ends of two of our spectra where the flux calibration uncertainties are large (and not reflected in the plotted uncertainties which are simply the statistical uncertainties on the flux measurement), and as such their fluxes are almost certainly underestimated given that they fall well below the trend of the other measured lines. P14 and P15 are not shown as they lie in a gap in our spectral coverage.

Temperatures determined from CEL ratios are around 10kK. The Balmer and Paschen jumps are well detected and permit the derivation of temperatures from their magnitudes. They are statistically consistent with each other, but slightly lower than the temperatures from CELs, with T(BJ)=8900800K, and T(PJ)=72001000K, where T([O iii])=10000200K. The ratios of the three strong helium lines at 4471, 5876 and 6678Å are weakly temperature-sensitive, and these are well enough detected to enable their use as a diagnostic; the temperatures from the 5876/4471 and the 6678/4471 ratios are significantly lower than from other aforementioned methods, at 3–4kK. Finally, O ii recombination line ratios imply a temperature consistent with the helium line ratios, albeit with a large statistical uncertainty, in addition to the systematic uncertainty of the effect of bleeding, which would result in an underestimate of the temperature.

Chemical abundances were derived from both collisionally excited lines and recombination lines (ORLs), with total chemical abundances derived using the ionisation correction scheme of Delgado-Inglada et al. (2014). Heavy element recombination lines are very well detected in the spectra, and as such it is possible to measure abundances for C, C, O, N, N and Ne. The neat code employed for the analysis assumes a three-zone ionization scheme. The average of ([O ii]) and ([S ii]) was adopted as representative of the density of low-ionisation zone (Ionisation Potential, IP 20 eV), while the average of ([Cl iii]) and ([Ar iv]) was applied for the medium ionisation zone (20 eV IP 45 eV). The electron temperature obtained from [N ii] was used for the temperature of the low-ionisation zone, and the average of the temperatures derived using [O iii], [Ar iii], and [S iii] lines taken as the temperature of the medium-ionisation zone. No lines of the high ionisation regime (IP45 eV) are detected, so no high-ionisation conditions are assumed. The medium-ionisation parameters were also employed in the derivation of abundances from ORLs. The calculated abundances, from both ORLs and CELs, are listed in Tables 2 and 3.

Ion
He (ORLs)
He (ORLs)
C (CELs)
C (ORLs)
C (ORLs)
N (CELs)
N (ORLs)
N (ORLs)
O (CELs)
O (ORLs)
O (CELs)
adf (O)
Ne (ORLs)
Ne (CELs)
adf (Ne)
Ar (CELs)
Ar (CELs)
S (CELs)
S (CELs)
Cl (CELs)
Cl (CELs)
Cl (CELs)
Table 2: Ionic abundance ratios, by number, relative to H, in IC 4776.

O permits the best determination of the abundance discrepancy, as ORLs and CELs of the ion are both present in optical spectra. 68 O ii recombination lines are detected, together with the three strong [O iii] CELs at 4363, 4959 and 5007Å. The abundances derived from the well-detected multiplets V1, V2, V10, V12, and V19 agree very well with each other; lines from multiplets V5, V20, V25, V28 and a number of 3d–4f transitions are detected, but imply much higher abundances and may be affected by noise or bleeding from strong lines in adjacent orders. As such, the total O abundance is derived based only on the 5 well detected multiplets. The abundance for each multiplet was determined from a flux-weighted average of the detected components, and the overall abundance calculated as the average of the multiplet abundances. The O/H abundance thus derived is (6.00.310, while that derived from the strong CELs is (3.40.210, giving an abundance discrepancy factor (adf; the ratio of ORL abundance to CEL abundance) of 1.750.15.

In the case of oxygen, O ORLs and CELs were detected while for O only CELs were detected. The He/H ratio implies that there should be negligible O, and so the CEL total abundance is simply the sum of O and O. For ORLs, it was assumed that the ionisation structure is the same, and that the O/O ratio from CELs can be used to correct for it. Thus, the abundance discrepancy factors for O and for O are the same.

Lines of Ne from both recombination and collisional excitation were also both detected in the spectra. The ORLs give abundances consistent with each other, yielding a value of (1.70.210. The CELs give (0.90.0510, resulting a discrepancy factor of 1.90.3, consistent with that derived for O.

Optical spectra contain CELs only of N, but ORLs of both N and N. Most of the N is in the form of N, and so the comparison of abundances relies on the large and uncertain correction of the CEL abundance for the unseen ions. Delgado-Inglada et al. (2014) proposed a new ICF scheme to compute N abundances; however, the differences with the classical ICF approach (N/O=N/O) can be very large, especially for relatively low-excitation PNe, as is the case of IC 4776. Using the newer ICF, we obtain N/H = , while the classical approach gives an N abundance 0.4 dex lower. Delgado-Inglada et al. (2015) found that the Delgado-Inglada et al. (2014) ICF for N showed a correlation with both He abundance and degree of ionisation and, as such, favoured the classical approach. As the ICF for N from Delgado-Inglada et al. (2014) is strongly dependent on whether the PN is radiation or matter-bounded (providing a more realistic correction for matter-bounded PNe; Delgado-Inglada, private communication), we used the classical ICF approach to derive N abundances for IC 4776. The derived abundance discrepancy factor for N is 2.9, with statistical uncertainties of , though with systematic uncertainties probably much larger (especially given the uncertainty in the choice of ICF). The net result is that, whatever ICF we chose, the N/O ratio is relatively low (-0.75 log (N/O) -0.35).

Without UV spectra, we cannot derive an abundance discrepancy factor for carbon. However, studies have generally found that ionic and elemental ratios from ORLs are consistent with those derived from CELs (Wang & Liu, 2007; Delgado-Inglada & Rodríguez, 2014). The C/O ratio derived from recombination lines is 0.3, atypically low for a planetary nebula (which normally present values around unity111 Wesson et al. (2005) found an average C/O ratio of 0.85 in a sample of nebulae outside the solar circle, while Kingsburgh & Barlow (1994) find a value of 1.15 in a sample interior to the solar circle - both measured from CELs.).

Very recently, Juan de Dios & Rodríguez (2017) have demonstrated the critical role of atomic data uncertainties in the determination of chemical abundances. They showed that the transition probabilities of the commonly used density diagnostic lines of S, O, Cl and Ar as well as the collision strengths of Ar are responsible for most of the uncertainty in the derived total abundances, especially in the high-density (log () ) regime. Given that the computed density for IC 4776 is relatively high, especially for the medium ionisation zone, it is important to note that the uncertainties derived for the total nebular abundances are almost certainly underestimated. However, the consistency between the adfs derived for O and Ne provides a strong indication of the validity of the results.

Figure 7: The alfa spectral fit (red-dashed) overlaid on top of the combined, observed UVES spectra (black-solid). The panels highlight key regions used in the neat analysis: (a) and (b) show two regions containing O ii recombination lines critical for deriving the nebular adf, while the panels (c) and (d) cover the Balmer and Paschen jumps, respectively. The full spectrum is shown in figure 10.
Element
He (ORLs)
C (ORLs)
N (ORLs)
N (CELs)
adf (N)
O (ORLs)
O (CELs)
adf (O)
Ne (ORLs)
Ne (CELs)
adf (Ne)
Ar (CELs)
S (CELs)
Cl (CELs)
Table 3: Total nebular abundances, relative to H, in IC 4776.

4 Central star

4.1 Observations and data reduction

The central star of IC 4776 was observed a total of 10 times using the FORS2 instrument mounted on ESO’s VLT-UT1 (Appenzeller et al., 1998). 1200s exposures were acquired in service mode, randomly distributed throughout the observing period (see Table 4 for the exact dates). For each observation, the same instrumental setup was used, employing a longslit of width 0.5″ and the GRIS_1200B grism resulting in a spectral range of 3700Å 5100Å with a spectral resolution of approximately 0.8Å.

All data were bias-subtracted, wavelength-calibrated and optimally-extracted using standard starlink routines (Shortridge et al., 2004).

4.2 Radial velocity variability

After the reduction, the spectra were continuum subtracted and aligned using the He i 4471.48 nebular line, such that the systemic nebular velocity was at 0 km s (thus accounting for both heliocentric velocity variations between observations and small deviations in the wavelength calibration). The Balmer lines were all found to be severely contaminated by the bright surrounding nebula, essentially leaving only the absorption line of He ii 4541 as a “clean” feature for cross-correlation. The individual spectra were all cross-correlated against a custom mask (namely a flat continuum with a deep, narrow absorption spike at the rest velocity of the He ii feature) with the resulting radial velocities (c.f. the nebular systemic velocity) listed in Table 4, and plotted in Figure 8.

Heliocentric Julian Radial velocity
Date (days) (km s)
2456428.84873 -20.53 1.80
2456444.70375 49.28 3.66
2456445.75583 17.03 5.28
2456447.81343 -4.98 1.42
2456460.80640 16.01 2.08
2456460.82896 26.72 4.59
2456486.59737 -11.03 4.61
2456508.64040 47.42 7.44
2456531.59275 -30.92 1.68
2456564.53724 -18.70 2.37
Table 4: Radial velocities of the central star of IC 4776 with respect to the nebular systemic velocity
Figure 8: Radial velocity curve of the central star of IC 4776 based on the FORS2 data with a tentative fit of period equal to 9 days. It is important to note that similarly good fits can be made with shorter periods, including periods around 0.3 and 1.2 days. Further observations are essential to fully constrain the period.

There is a clear variation in the radial velocity with a semi-amplitude of about 30–40 km s strongly indicating that the central star of IC 4776 is a binary system. Unfortunately, with so few radial velocity points, it is not possible to fully constrain the orbital period of the likely binary, with several periodicities and amplitudes providing reasonable fits to the data. Our current radial velocity data rule out periods much longer than 20 days, but shorter periods are relatively well fit. In Figure 8, we show a tentative fit with a period of 9 days to highlight the significant radial velocity variations most likely due to orbital motion, as well as the need for further data. Interestingly, although the stellar systemic velocity (the zero of the sine curve used to fit the stellar radial velocities, often referred to as ) was allowed to vary as part of the fitting procedure, it was found to coincide with the systemic velocity of the nebula (at zero in the plot given that the all radial velocities have been measured with respect to the nebular systemic velocity) within uncertainties (3 km s). This indicates that the plotted fit may indeed be close to the true orbital solution. However, it must be emphasised that we find similarly good fits for a range of periods, including 0.3 and 1.2 days.

It is important to note that we do not see any clear signs of a secondary component in our spectra. Previous authors have classified the central star as a wels type (Tylenda et al., 1993), indicating that perhaps some emission lines attributable to the secondary could be present in the spectra. However, the extremely bright and compact nature of the central region of the nebula can account for the presence of these lines (Basurah et al., 2016). We find the same bright, narrow emission lines in our 2D spectra, but find no evidence of an appreciable contribution to their flux from the central star; rather, they originate purely from the compact nebula (although accurate nebular subtraction is a challenge). Assuming that no such stellar emission lines are present and that the secondary is a main sequence star (the most common scenario; Jones & Boffin, 2017), short periods (1 day) are unlikely given that at short periods a main-sequence star would be expected to show strong emission lines due to irradiation (Miszalski et al., 2011b; Jones et al., 2014; Jones et al., 2015). However, if the secondary is a fainter white dwarf, then short periods cannot be ruled out on the basis of a non-detection of irradiated lines in the spectrum.

Assuming that the nebular inclination derived in Section 2.3 is reflective of the inclination of the binary (as found in all cases where both inclinations are known; Hillwig et al., 2016), then it is possible to place some limits on the possible mass of the secondary. Further assuming that the amplitude of the fit shown in Figure 8, K40 km s, is representative of the true amplitude (a seemingly reasonable assumption given the relatively even distribution of the radial velocity data points), then a 9 day period would imply a mass ratio (=) of around unity. Longer periods and/or greater radial velocity amplitudes would both imply higher mass ratios and vice versa. The mass function for the secondary can be written as

(1)

where and are the primary and secondary masses, is the inclination of the orbital plane, and is the orbital period.

Taking a relatively standard white dwarf mass for the primary of M=0.6M (and maintaining the assumption that K40 km s), the mass function can be solved analytically for the secondary mass. Taking into account the uncertainty on the nebular inclination (based on the spatiokinematical modelling presented in Section 2.3), the possible range of secondary masses are plotted as a function of period in Figure 9. Under these assumptions, the tentative 9 day period would imply that the secondary is likely a late-type main sequence star (roughly of spectral type K) or a white dwarf. For the secondary to be a white dwarf, it would have to have been initially the more massive component of the binary and, as such, would be expected to leave a more massive remnant (at least more massive than the primary remnant which, in this case, was assumed to be 0.6M). This is a possibility given that the open mass range for the secondary, based on our fits, is approximately 0.6–0.7M. However, a more massive secondary would probably be required in order for it to have now cooled beyond observability in our spectra. If the true period is shorter than 9 days (Figure 9 also shows solutions for the reasonable fits at 0.3 days and 1.2 days) and then the open mass range would be significantly lower, ruling out an evolved white dwarf companion for the same reason, while a main sequence secondary would become less likely given the aforementioned non-detection of irradiated emission lines originating from the secondary. However, this effect may be counteracted by the implied low secondary masses (and therefore radii) at shorter periods, as the level of irradiation in such systems is principally a function of the apparent radius of the irradiated star (proportional to the ratio of the secondary radius and the orbital separation).

Greater radial velocity amplitudes than the assumed 40 km s would open the possibility of a more massive white dwarf companion, but the currently available data provides no indication of such a large amplitude. Similarly, periods longer than 9 days open the parameter space to the possibility of a more massive, evolved white dwarf companion, however in that case the system would have to have undergone two common-envelope phases (given that the AGB radius of a more massive companion would inevitably have been larger than that of the primary) which almost certainly would result in a very short orbital period (Tovmassian et al., 2010). Therefore, on a stellar evolutionary basis, a main sequence secondary (roughly of K-type for a period of 9 day, or late M-type for periods less than 1 day) would seem the more plausible companion, although a white dwarf secondary cannot be entirely ruled out.

Figure 9: A plot showing the possible secondary masses of IC 4776 (shaded region) as a function of period for the FORS2 radial velocity data. The plot assumes the same radial velocity amplitude as the fit of Figure 8 and a 0.6M primary in a circular orbit. The range of possible masses for a given period is derived assuming that the binary is coplanar with the waist of the nebula, and the range of values for a given period is a reflection of the uncertainty on the nebular inclination. The periods which offer similarly good fits to the radial velocity data are delineated to highlight the approximate range of possible secondary masses ( M) For more detail refer to the text.

5 Discussion

Population synthesis models predict many post-CE systems with periods of several to a few tens of days (e.g. Han et al., 1995). However, the known population of post-CE binary central stars of PNe is very sparse in this region (De Marco et al., 2008; Jones & Boffin, 2017), with a similar paucity observed in the general white dwarf plus main sequence binary population (Nebot Gómez-Morán et al., 2011). Thus far, it is unclear whether this lack of intermediate period binary central stars represents purely failure of the population synthesis models222It is important to note that population synthesis models rely on an ad hoc prescription of the common envelope process, often relying on simple parameterisations that have been proven to be less than satisfactory in reproducing observations (De Marco et al., 2011). Hydrodynamic models thus far fall short of being able to make the predictions required for use in population synthesis (Ivanova et al., 2013), however tend to almost unanimously predict very short post-CE orbitals periods while the few that do predict longer periods seem to be under-resolved (see e.g. Passy et al., 2012). or also an observational bias towards the discovery of short period systems.

Recent observations have indicated that long-term radial velocity monitoring with modern, high-stability spectrographs may hold the key to revealing this missing intermediate-period population, should it indeed exist (e.g. Miszalski et al., 2017). Here, we have reported on the discovery of a post-CE binary star at the centre of IC 4776 via such radial velocity monitoring. While the period of the system cannot be confirmed by the data (several periods ranging from 0.3–9 days provide reasonable orbital solutions), the detection further supports the hypothesis that a systematic radial velocity survey may indeed reveal the presence of many intermediate period binaries inside PNe.

The central star of IC 4776 was found to be a single-lined binary with a tentative orbital period of 9 days (though shorter periods cannot be ruled out), where the secondary is most likely a main sequence star (of mass 0.1–0.7 M). The system was selected for monitoring based primarily on the presence of precessing jets, believed to result from mass transfer in a central binary system. Such mass transfer has been hypothesized to occur either prior to or immediately after the CE phase (Corradi et al., 2011; Miszalski et al., 2013; Boffin et al., 2012; Tocknell et al., 2014), but may also help to prevent in-spiral during the CE as part of a grazing envelope evolution (GEE; Soker, 2015; Shiber et al., 2017). Such a GEE may lead to the preferential formation of wide binaries, but the range of possible periods derived for IC 4776 does not extend to long enough periods to necessitate a GEE to explain its formation (though nor does it rule out such an evolution).

An empirical analysis of the chemical abundances in IC 4776 based on échelle spectra covering the entire optical range indicates a low N/O ratio as well as a low C/O ratio (both of which are 0.3). This may be indicative that interaction with the companion cut short the AGB evolution of the nebular progenitor (De Marco, 2009; Jones et al., 2014).

In PNe, there is a long standing discrepancy between chemical abundances derived using ORLs and those derived using CELs, which has often been attributed to the existence of a second phase of cold hydrogen deficient material within the normal nebular gas phase (Liu et al., 2000; Tsamis et al., 2003; Zhang et al., 2004; Wesson et al., 2005). The discrepancy is typically a factor of 2–3 but exceeds a factor of 5 in about 20% of nebulae. Recent observations have suggested that high abundance discrepancies may be strongly correlated with central stars that have undergone a CE evolution (Corradi et al., 2015; Jones et al., 2016; Wesson et al., 2017). However, our chemical analysis reveals the adf of IC 4776 to be particularly low at 2, which is not just low for a PN with a binary central star, but for PNe in general. Interestingly, the only other post-CE system known to reside in a low-adf nebula is an intermediate-period system (NGC 5189 with a period of 4 days; García-Rojas et al., 2013; Manick et al., 2015) which, if the period of IC 4776 is a long as 9 days, may imply a connection between intermediate period binaries and low nebular adfs. Furthermore, NGC 5189 is host to a [WR] type primary (Manick et al., 2015), as has been claimed for the central star of IC 4776 (Tylenda et al., 1993). In total, this means three [WR]-type central stars are known to reside in binaries, with the two previous detections (NGC 5189 and PM 1-23) both having orbital periods longer than one day333The orbital period of the [WR] central star of PM 1-23 is given as 0.6 days by Hajduk et al. (2010), but further observations have shown that this is, in fact, half the orbital period (Manick et al., 2015, Miszalski et al., in prep.)..

The connection between intermediate periods, low adfs and [WR] central stars is, thus far, tenuous, given the extremely small number statistics involved, particularly given that we are unable to derive a definitive period for the central star of IC 4776. However, there may perhaps be reasonable physical grounds for such correlations. Manick et al. (2015) suggest that the strong winds observed in [WR] stars may help to prevent spiral-in, via pre-CE wind interaction, leading to longer post-CE periods444A similar scenario has been suggested for the intermediate-period binary central star of NGC 2346 (V651 Mon, the period of which is 16 days; Soker, 2002), but there is no indication that its central star is of the [WR]-type..

Should high adfs be the result of a nova-like eruption which ejects a second, chemically-enriched gas phase into the nebula (Corradi et al., 2015; Jones et al., 2016; García-Rojas et al., 2016), then it is not unreasonable to expect that central stars with little hydrogen on their surfaces may be less likely to experience such outbursts (i.e. without hydrogen to burn, they do not undergo an eruption). However, one formation mechanism for such [WR] stars is to undergo some form of late thermal pulse which depletes their outer layers in hydrogen, whilst also providing an explanation for the presence of hydrogen-poor ejecta in their host nebulae (Ercolano et al., 2004). Indeed, previously connections were made between the abundance discrepancy problem and hydrogen-depleted central stars (Ercolano et al., 2004), particularly in the case of born-again central stars which are found to show some of the highest discrepancies (Wesson et al., 2003; Wesson et al., 2008). However, more recent studies have shown that in many cases PNe with [WR] central stars present with rather modest abundance discrepancies (García-Rojas et al., 2013).

Assuming that high adfs are the result of late ejection of low-metallicity material, it might be expected that longer period systems do not experience any post-CE mass transfer that could lead to an outburst. There is little evidence of such post-CE mass transfer in other high adf PNe, however, in NGC 6337, jets are observed to have been launched after the ejection of the CE (perhaps as a result of post-CE mass transfer; García-Díaz et al., 2009) while the nebular adf is observed to be highly elevated (30; Wesson et al., 2017). Furthermore, the possibility remains open for other systems based on the observation of near-Roche-lobe-filling, post-CE secondary stars (Jones et al., 2015).

Unfortunately, the adf of the only other PN known to host a [WR]-binary central star (PM 1-23) is not known, nor are the adfs of other PNe known to host long- and intermediate-period post-CE central stars (e.g. LTNF 1, Sp 1, 2MASS J19310888+4324577, MPA J15086455, NGC 2346, NGC 1360). Thus, making further assessment of the possible relationships between periodicity, abundance pattern and central star type currently impossible.

In conclusion, in spite of the laborious nature of the observations (De Marco et al., 2004), we encourage further radial velocity studies of the central stars of planetary nebulae in search of the missing population of intermediate-period post-CE binaries. As highlighted here, the pre-selection of targets based on morphology may be the most effective means of increasing the hit-rate for detections. Whilst the possibility of a connection between [WR]-type central stars and intermediate periods is intriguing (as well as a possible link between [WR]-type and low adfs), a more rigorous survey of [WR] central stars (and their host nebulae) is required before any strong conclusions can be drawn. Furthermore, a connection between intermediate-period central stars and low adfs is on a similarly uncertain footing, and abundance studies of the known sample of PNe known to host binary central stars (of all periodicities) is strongly encouraged. Further observations of the central star of IC 4776 are essential in order to full constrain its period, and begin to place these nascent relationships on a stronger footing.

Acknowledgements

We are grateful to the referee, Orsola De Marco, for helpful comments and suggestions that further improved the paper. We would like to thank Tom Marsh for the use of his molly software package, and Aníbal García-Hernandez for the use of his UVES data. We thank Gloria Delgado-Inglada for fruitful discussions. Based on data obtained at the European Southern Observatory, Chile, under proposal numbers 097.D-0037, 091.D-0673, 089.D-0429, 087.D-0446. This research made use of Astropy, a community-developed core Python package for Astronomy (Astropy Collaboration et al., 2013). PS acknowledges the ING Support and Research Studentship, and thanks the Polish National Center for Science (NCN) for support through grant 2015/18/A/ST9/00578. J. G-R acknowledges support from Severo Ochoa Excellence Program (SEV-2015-0548) Advanced Postdoctoral Fellowship.

References

  • Aller & Keyes (1985) Aller L. H., Keyes C. D., 1985, PASP, 97, 1142
  • Appenzeller et al. (1998) Appenzeller I., et al., 1998, The Messenger, 94, 1
  • Astropy Collaboration et al. (2013) Astropy Collaboration et al., 2013, A&A, 558, A33
  • Basurah et al. (2016) Basurah H. M., Ali A., Dopita M. A., Alsulami R., Amer M. A., Alruhaili A., 2016, MNRAS, 458, 2694
  • Boffin et al. (2012) Boffin H. M. J., Miszalski B., Rauch T., Jones D., Corradi R. L. M., Napiwotzki R., Day-Jones A. C., Köppen J., 2012, Science, 338, 773
  • Bryce et al. (1997) Bryce M., Lopez J. A., Holloway A. J., Meaburn J., 1997, ApJ Letters, 487, 161
  • Corradi & Schwarz (1993) Corradi R. L. M., Schwarz H. E., 1993, A&A, 268, 714
  • Corradi et al. (2011) Corradi R. L. M., et al., 2011, MNRAS, 410, 1349
  • Corradi et al. (2015) Corradi R. L. M., García-Rojas J., Jones D., Rodríguez-Gil P., 2015, ApJ, 803, 99
  • De Marco (2009) De Marco O., 2009, PASP, 121, 316
  • De Marco & Soker (2011) De Marco O., Soker N., 2011, PASP, 123, 402
  • De Marco et al. (2004) De Marco O., Bond H. E., Harmer D., Fleming A. J., 2004, ApJ, 602, 93
  • De Marco et al. (2008) De Marco O., Hillwig T. C., Smith A. J., 2008, AJ, 136, 323
  • De Marco et al. (2011) De Marco O., Passy J.-C., Moe M., Herwig F., Mac Low M.-M., Paxton B., 2011, MNRAS, 411, 2277
  • Dekker et al. (2000) Dekker H., D’Odorico S., Kaufer A., Delabre B., Kotzlowski H., 2000, in M. Iye & A. F. Moorwood ed., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series Vol. 4008, Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series. pp 534–545
  • Delgado-Inglada & Rodríguez (2014) Delgado-Inglada G., Rodríguez M., 2014, ApJ, 784, 173
  • Delgado-Inglada et al. (2014) Delgado-Inglada G., Morisset C., Stasińska G., 2014, MNRAS, 440, 536
  • Delgado-Inglada et al. (2015) Delgado-Inglada G., Rodríguez M., Peimbert M., Stasińska G., Morisset C., 2015, MNRAS, 449, 1797
  • Durand et al. (1998) Durand S., Acker A., Zijlstra A., 1998, A&AS, 132, 13
  • Ercolano et al. (2004) Ercolano B., Wesson R., Zhang Y., Barlow M. J., De Marco O., Rauch T., Liu X.-W., 2004, MNRAS, 354, 558
  • Frew et al. (2016) Frew D. J., Parker Q. A., Bojičić I. S., 2016, MNRAS, 455, 1459
  • García-Díaz et al. (2009) García-Díaz M. T., Clark D. M., López J. A., Steffen W., Richer M. G., 2009, ApJ, 699, 1633
  • García-Rojas et al. (2013) García-Rojas J., Peña M., Morisset C., Delgado-Inglada G., Mesa-Delgado A., Ruiz M. T., 2013, A&A, 558, A122
  • García-Rojas et al. (2016) García-Rojas J., Corradi R. L. M., Monteiro H., Jones D., Rodríguez-Gil P., Cabrera-Lavers A., 2016, ApJL, 824, L27
  • García-Segura et al. (2014) García-Segura G., Villaver E., Langer N., Yoon S.-C., Manchado A., 2014, ApJ, 783, 74
  • Górny et al. (2010) Górny S. K., Perea-Calderón J. V., García-Hernández D. A., García-Lario P., Szczerba R., 2010, A&A, 516, A39
  • Guzman-Ramirez et al. (2014) Guzman-Ramirez L., Lagadec E., Jones D., Zijlstra A. A., Gesicki K., 2014, MNRAS, 441, 364
  • Hajduk et al. (2010) Hajduk M., Zijlstra A. A., Gesicki K., 2010, MNRAS, 406, 626
  • Han et al. (1995) Han Z., Podsiadlowski P., Eggleton P. P., 1995, MNRAS, 272, 800
  • Hillwig et al. (2016) Hillwig T. C., Jones D., Marco O. D., Bond H. E., Margheim S., Frew D., 2016, The Astrophysical Journal, 832, 125
  • Howarth (1983) Howarth I. D., 1983, MNRAS, 203, 301
  • Ivanova et al. (2013) Ivanova N., et al., 2013, A&ARv, 21, 59
  • Jones & Boffin (2017) Jones D., Boffin H. M. J., 2017, Nature Astronomy, 1, 0117
  • Jones et al. (2014) Jones D., Boffin H. M. J., Miszalski B., Wesson R., Corradi R. L. M., Tyndall A. A., 2014, A&A, 562, A89
  • Jones et al. (2015) Jones D., Boffin H. M. J., Rodríguez-Gil P., Wesson R., Corradi R. L. M., Miszalski B., Mohamed S., 2015, A&A, 580, A19
  • Jones et al. (2016) Jones D., Wesson R., García-Rojas J., Corradi R. L. M., Boffin H. M. J., 2016, MNRAS, 455, 3263
  • Jones et al. (2017) Jones D., Van Winckel H., Aller A., Exter K., De Marco O., 2017, A&A, 600, L9
  • Juan de Dios & Rodríguez (2017) Juan de Dios L., Rodríguez M., 2017, preprint, (arXiv:1704.06009)
  • Kahn & West (1985) Kahn F. D., West K. A., 1985, MNRAS, 212, 837
  • Kingsburgh & Barlow (1994) Kingsburgh R. L., Barlow M. J., 1994, MNRAS, 271, 257
  • Kwok et al. (1978) Kwok S., Purton C. R., Fitzgerald P. M., 1978, ApJ, 219, 125
  • Liu et al. (2000) Liu X.-W., Storey P. J., Barlow M. J., Danziger I. J., Cohen M., Bryce M., 2000, MNRAS, 312, 585
  • Manick et al. (2015) Manick R., Miszalski B., McBride V., 2015, MNRAS, 448, 1789
  • Méndez & Niemela (1981) Méndez R. H., Niemela V. S., 1981, ApJ, 250, 240
  • Miszalski et al. (2009a) Miszalski B., Acker A., Moffat A. F. J., Parker Q. A., Udalski A., 2009a, A&A, 496, 813
  • Miszalski et al. (2009b) Miszalski B., Acker A., Parker Q. A., Moffat A. F. J., 2009b, A&A, 505, 249
  • Miszalski et al. (2011a) Miszalski B., Corradi R. L. M., Boffin H. M. J., Jones D., Sabin L., Santander-García M., Rodríguez-Gil P., Rubio-Díez M. M., 2011a, MNRAS, 413, 1264
  • Miszalski et al. (2011b) Miszalski B., Jones D., Rodríguez-Gil P., Boffin H. M. J., Corradi R. L. M., Santander-García M., 2011b, A&A, 531, A158
  • Miszalski et al. (2013) Miszalski B., Boffin H. M. J., Corradi R. L. M., 2013, MNRAS, 428, L39
  • Miszalski et al. (2017) Miszalski B., Manick R., Mikołajewska J., Iłkiewicz K., Kamath D., Van Winckel H., 2017, preprint, (arXiv:1703.10891)
  • Nebot Gómez-Morán et al. (2011) Nebot Gómez-Morán A., et al., 2011, A&A, 536, A43
  • Nordhaus et al. (2007) Nordhaus J., Blackman E. G., Frank A., 2007, MNRAS, 376, 599
  • O’Connor et al. (2000) O’Connor J. A., Redman M. P., Holloway A. J., Bryce M., López J. A., Meaburn J., 2000, ApJ, 531, 336
  • Parker et al. (2006) Parker Q. A., et al., 2006, MNRAS, 373, 79
  • Pasquini et al. (2002) Pasquini L., et al., 2002, The Messenger, 110, 1
  • Passy et al. (2012) Passy J.-C., et al., 2012, ApJ, 744, 52
  • Perea-Calderón et al. (2009) Perea-Calderón J. V., García-Hernández D. A., García-Lario P., Szczerba R., Bobrowsky M., 2009, A&A, 495, L5
  • Phillips & Pottasch (1984) Phillips J. P., Pottasch S. R., 1984, A&A, 130, 91
  • Raga et al. (2009) Raga A. C., Esquivel A., Velázquez P. F., Cantó J., Haro-Corzo S., Riera A., Rodríguez-González A., 2009, ApJL, 707, L6
  • Shiber et al. (2017) Shiber S., Kashi A., Soker N., 2017, MNRAS, 465, L54
  • Shortridge et al. (2004) Shortridge K., et al., 2004, Starlink User Note 86.21. Rutherford Appleton Laboratory
  • Soker (2002) Soker N., 2002, MNRAS, 336, 1229
  • Soker (2006) Soker N., 2006, PASP, 118, 260
  • Soker (2015) Soker N., 2015, ApJ, 800, 114
  • Stanghellini et al. (2008) Stanghellini L., Shaw R. A., Villaver E., 2008, ApJ, 689, 194
  • Steffen et al. (2009) Steffen W., García-Segura G., Koning N., 2009, ApJ, 691, 696
  • Steffen et al. (2010) Steffen W., Koning N., Wenger S., Morisset C., Magnor M., 2010, IEEE Transactions on Visual and Computer Graphics
  • Tocknell et al. (2014) Tocknell J., De Marco O., Wardle M., 2014, MNRAS, 439, 2014
  • Toonen & Nelemans (2013) Toonen S., Nelemans G., 2013, A&A, 557, A87
  • Tovmassian et al. (2010) Tovmassian G., et al., 2010, ApJ, 714, 178
  • Tsamis et al. (2003) Tsamis Y. G., Barlow M. J., Liu X.-W., Danziger I. J., Storey P. J., 2003, MNRAS, 345, 186
  • Tylenda et al. (1993) Tylenda R., Acker A., Stenholm B., 1993, A&AS, 102, 595
  • Wang & Liu (2007) Wang W., Liu X.-W., 2007, MNRAS, 381, 669
  • Wesson (2016) Wesson R., 2016, MNRAS, 456, 3774
  • Wesson et al. (2003) Wesson R., Liu X.-W., Barlow M. J., 2003, MNRAS, 340, 253
  • Wesson et al. (2005) Wesson R., Liu X.-W., Barlow M. J., 2005, MNRAS, 362, 424
  • Wesson et al. (2008) Wesson R., Barlow M. J., Liu X.-W., Storey P. J., Ercolano B., De Marco O., 2008, MNRAS, 383, 1639
  • Wesson et al. (2012) Wesson R., Stock D. J., Scicluna P., 2012, MNRAS, 422, 3516
  • Wesson et al. (2017) Wesson R., Jones D., García-Rojas J., Corradi R. L. M., Boffin H. M. J., 2017, preprint, (arXiv:1612.02215)
  • Zhang et al. (2004) Zhang Y., Liu X.-W., Wesson R., Storey P. J., Liu Y., Danziger I. J., 2004, MNRAS, 351, 935

Appendix A UVES spectrum of IC 4776

Figure 10 shows the full UVES spectrum of IC 4776 referred to in Section 2, along with the alfa fit over-plotted. The observed and dereddened fluxes derived from the alfa fit are presented in Table 5.

(Å) Ion Multiplet Lower term Upper term g g
3312.32 0.037 0.007 0.029 0.008 iii V3 3s 3P* 3p 3S 3 3
3322.53 0.050 0.015 0.071 0.018 [Fe iii] 5D 7S 9 7
3334.84 0.100 0.012 0.122 0.015 Ne ii V2 3s 4P 3p 4D* 6 8
3342.50 0.171 0.020 0.173 [Ne iii] 2p4 1D 2p4 1S 5 1
3353.17 0.036 0.011 0.060 0.013 [Cl iii] 3p3 4S* 3p3 2P* 4 2
3354.55 0.095 0.014 0.127 0.017 He i V8 2s 1S 7p 1P* 1 3
3355.02 0.035 0.014 0.060 Ne ii V2 3s 4P 3p 4D* 4 6
3417.69 0.024 0.005 0.033 0.005 Ne ii V19 3p 2D* 3d 4F 6 8
3444.07 0.034 0.007 0.029 0.008 iii V15 3p 3P 3d 3P* 5 5
3447.59 0.155 0.015 0.216 He i V7 2s 1S 6p 1P* 1 3
3453.32 0.013 0.008 0.034 0.009 He i V48 2p 3P* 20d 3D 9 15
3456.86 0.054 0.007 0.052 He i V47 2p 3P* 19d 3D 9 15
3461.01 0.049 0.006 0.052 He i V46 2p 3P* 18d 3D 9 15
3465.94 0.047 0.011 0.050 0.013 He i V45 2p 3P* 17d 2D 9 15
3471.81 0.076 0.008 0.063 0.010 He i V44 2p 3P* 16d 3D 9 15
3478.97 0.061 0.009 0.067 0.011 He i V43 2p 3P* 15d 3D 9 15
3487.72 0.082 0.009 0.095 0.011 He i V42 2p 3P* 14d 3D 9 15
3498.64 0.115 0.013 0.111 0.016 He i V40 2p 3P* 13d 3D 9 15
3512.52 0.117 0.012 0.132 He i V38 2p 3P* 12d 3D 9 15
3530.50 0.157 0.012 0.165 He i V36 2p 3P* 11d 3D 9 15
3554.42 0.203 0.013 0.226 He i V34 2p 3P* 10d 3D 9 15
3568.50 0.016 0.004 0.021 0.005 Ne ii V9 3s’ 2D 3p’ 2F* 6 8
3587.28 0.302 0.018 0.316 He i V31 2p 3P* 9d 3D 9 15
3613.64 0.329 0.017 0.369 0.021 He i V6 2s 1S 5p 1P* 1 3
3634.25 0.421 0.021 0.473 He i V28 2p 3P* 8d 3D 9 15
3662.26 0.182 0.046 0.187 0.054 i H30 2p+ 3P* 30d+ 2D 8 *
3663.40 0.172 0.019 0.236 i H29 2p+ 3P* 29d+ 2D 8 *
3664.68 0.275 0.016 0.301 i H28 2p+ 3P* 28d+ 2D 8 *
3666.10 0.308 0.018 0.334 0.022 i H27 2p+ 2P* 27d+ 2D 8 *
3667.68 0.346 0.021 0.414 i H26 2p+ 2P* 26d+ 2D 8 *
3669.46 0.384 0.024 0.416 i H25 2p+ 3P* 25d+ 2D 8 *
3671.48 0.489 0.024 0.502 i H24 2p+ 2P* 24d+ 2D 8 *
3673.74 0.539 0.028 0.597 i H23 2p+ 2P* 23d+ 2D 8 *
3676.36 0.558 0.028 0.630 0.035 i H22 2p+ 2P* 22d+ 2D 8 *
3679.36 0.594 0.034 0.713 i H21 2p+ 2P* 21d+ 2D 8 *
3682.81 0.672 0.039 0.776 i H20 2p+ 2P* 20d+ 2D 8 *
3686.83 0.750 0.041 0.881 0.051 i H19 2p+ 2P* 19d+ 2D 8 *
3691.56 0.955 0.048 0.996 i H18 2p+ 2P* 18d+ 2D 8 *
3697.15 1.023 0.054 1.117 i H17 2p+ 2P* 17d+ 2D 8 *
3703.86 1.217 0.057 1.314 0.070 i H16 2p+ 2P* 16d+ 2D 8 *
3705.02 0.629 0.066 0.729 He i V25 2p 3P* 7d 3D 9 15
3711.97 1.374 0.070 1.534 i H15 2p+ 2P* 15d+ 2D 8 *
3715.08 0.014 0.004 0.015 0.004 iii V14 3p 3P 3d 3D* 5 7
3721.63 1.206 0.091 1.192 [S iii] F2 3p2 3P 3p2 1S 3 1
3721.94 1.788 0.092 2.120 0.112 i H14 2p+ 2P* 14d+ 2D 8 *
3726.03 19.062 1.098 22.400 1.300 [O ii] F1 2p3 4S* 2p3 2D* 4 4
3728.82 8.352 0.660 9.471 [O ii] F1 2p3 4S* 2p3 2D* 4 6
3734.37 1.754 0.085 2.119 i H13 2p+ 2P* 13d+ 2D 8 *
3750.15 2.524 0.113 2.719 0.140 i H12 2p+ 2P* 12d+ 2D 8 *
3754.69 0.013 0.004 0.015 0.001 iii V2 3s 3P* 3p 3D 3 5
3754.70 * * iii V2 3s 3P* 3p 3D 3 5
3759.87 0.013 0.003 0.020 0.003 iii V2 3s 3P* 3p 3D 5 7
3770.63 2.984 0.145 3.479 0.175 i H11 2p+ 2P* 11d+ 2D 8 *
3777.14 0.009 0.004 0.015 0.005 Ne ii V1 3s 4P 3p 4P* 2 4
3784.89 0.041 0.003 0.042 He i ?
3791.27 0.012 0.003 0.011 0.003 iii V2 3s 3P* 3p 3D 5 5
3797.90 4.595 0.185 4.943 i H10 2p+ 2P* 10d+ 2D 8 *
3805.74 0.057 0.008 0.059 0.009 He i V58 2p 1P* 11d 1D 3 5
3819.62 1.180 0.030 1.254 0.041 He i V22 2p 3P* 6d 3D 9 15
3645.50 0.556 0.008 0.644 Balmer Jump-
Table 5: Observed, , and dereddened, , nebular emission line fluxes of IC 4776 relative to (). Asterisks denote that the line is detected as a blend with the previous line in the table, and the flux listed is the total flux of the unresolved lines. The upper and lower term columns show the upper and lower levels of the atomic transition producing the line, while g and g are the degeneracy factors of the two levels.

Observed and dereddened emission line fluxes from IC 4776 (Å) Ion Multiplet Lower term Upper term g g 3646.50 0.195 0.006 0.222 0.008 Balmer Jump+ 3819.62 1.154 0.027 1.323 0.040 He i V22 2p 3P* 6d 3D 9 15 3835.39 6.048 0.066 6.865 i H9 2p+ 2P* 9d+ 2D 8 * 3856.02 0.082 0.008 0.093 0.002 iii V12 3p2 2D 4p 2P* 6 4 3856.13 * * ii V12 3p 4D* 3d 4D 4 2 3862.60 0.064 0.005 0.079 Si ii V1 3p2 2D 4p 2P* 4 2 3868.75 70.487 2.963 77.200 3.600 [Ne iii] F1 2p4 3P 2p4 1D 5 5 3882.19 0.018 0.004 0.017 ii V12 3p 4D* 3d 4D 8 8 3888.65 4.196 0.368 4.954 He i V2 2s 3S 3p 3P* 3 9 3889.05 8.431 0.369 9.911 0.449 i H8 2p+ 2P* 8d+ 2D 8 * 3907.46 0.011 0.003 0.010 0.003 ii V11 3p 4D* 3d 4P 6 6 3918.98 0.011 0.002 0.013 0.002 ii V4 3p 2P* 4s 2S 2 2 3920.69 0.012 0.006 0.026 0.007 ii V4 3p 2P* 4s 2S 4 2 3926.54 0.131 0.006 0.140 0.007 He i V58 2p 1P* 8d 1D 3 5 3964.73 0.840 0.022 0.973 0.029 He i V5 2s 1S 4p 1P* 1 3 3967.46 22.782 0.930 25.300 1.100 [Ne iii] F1 2p4 3P 2p4 1D 3 5 3970.07 13.802 0.234 15.258 0.357 i H7 2p+ 2P* 7d+ 2D 8 98 3994.99 0.015 0.003 0.014 0.004 ii V12 3s 1P* 3p 1D 3 5 4009.26 0.215 0.018 0.188 He i V55 2p 1P* 7d 1D 3 5 4026.08 1.933 0.036 2.142 ii V39b 3d 3F* 4f 2[5] 7 9 4026.21 * * He i V18 2p 3P* 5d 3D 9 15 4041.31 0.012 0.001 0.012 0.001 ii V39b 3d 3F* 4f 2[5] 9 11 4043.53 0.006 0.002 0.007 0.002 ii V39a 3d 3F* 4f 2[4] 7 9 4060.60 0.010 0.003 0.012 0.003 ii V97 3d 2F 4f 2[4]* 8 * 4068.60 2.199 0.147 2.394 0.165 [S ii] F1 2p3 4S* 2p3 2P* 4 4 4072.16 0.115 0.008 0.125 ii V10 3p 4D* 3d 4F 6 8 4076.35 0.857 0.045 0.848 0.051 [S ii] F1 2p3 4S* 2p3 2P* 2 4 4078.84 0.035 0.008 0.045 0.009 ii V10 3p 4D* 3d 4F 4 4 4089.29 0.074 0.013 0.088 0.014 ii V48a 3d 4F 4f G5* 10 12 4092.93 0.014 0.004 0.015 0.004 ii V10 3p 4D* 3d 4F 8 8 4097.25 0.129 0.010 0.142 0.002 ii V48b 3d 4F 4f G4* 8 10 4097.26 * * ii V48b 3d 4F 4f G4* 8 10 4097.33 * * iii V1 3s 2S 3p 2P* 2 4 4101.74 24.309 0.357 26.903 0.536 i H6 2p+ 2P* 6d+ 2D 8 72 4110.78 0.028 0.002 0.029 ii V20 3p 4P* 3d 4D 4 2 4119.22 0.062 0.003 0.066 0.003 ii V20 3p 4P* 3d 4D 6 8 4120.54 0.044 0.004 0.051 0.004 ii V20 3p 4P* 3d 4D 6 4 4120.84 0.256 0.004 0.275 0.006 He i V16 2p 3P* 5s 3S 9 3 4121.46 0.024 0.003 0.027 0.003 ii V19 3p 4P* 3d 4P 2 2 4132.80 0.040 0.004 0.046 0.004 ii V19 3p 4P* 3d 4P 2 4 4143.76 0.319 0.008 0.347 0.010 He i V53 2p 1P* 6d 1D 3 5 4146.00 0.023 0.003 0.027 0.003 ii V106 3p 6P 3d 6D* 8 8 4153.30 0.045 0.002 0.053 0.002 ii V19 3p 4P* 3d 4P 4 6 4168.97 0.032 0.001 0.033 0.002 He i V52 2p 1P* 6s 1S 3 1 4169.22 0.020 0.001 0.022 0.001 ii V19 3p 4P* 3d 4P 6 6 4185.45 0.039 0.002 0.038 0.003 ii V36 3p’ 2F* 3d’ 2G 6 8 4189.79 0.043 0.003 0.047 0.003 ii V36 3p’ 2F* 3d’ 2G 8 10 4219.74 0.015 0.002 0.015 0.002 Ne ii V52a 3d 4D 4f 2[4]* 8 10 4241.78 0.010 0.002 0.010 0.002 ii V48b 3d 3D* 4f 1[4] 7 9 4242.49 0.009 0.002 0.008 0.002 ii V48b 3d 3D* 4f 1<4> 7 7 4243.98 0.061 0.005 0.061 [Fe ii] 4253.86 0.027 0.003 0.029 0.001 ii V101 3d 2G 4f 2[5]* 10 10 4254.00 * * ii V101 3d’ 2G 4f’ H5* 18 22 4267.15 0.148 0.002 0.164 0.003 ii V6 3d 2D 4f 2F* 10 14 4275.55 0.030 0.002 0.032 0.002 ii V67a 3d 4D 4f F4* 8 10 4276.28 0.013 0.002 0.015 0.003 ii V67b 3d 4D 4f F3* 6 6 4276.75 0.051 0.002 0.054 0.002 ii V67b 3d 4D 4f F3* 6 8 4277.43 0.010 0.002 0.008 0.002 ii V67c 3d 4D 4f F2* 2 4 4277.89 0.003 0.001 0.005 ii V67b 3d 4D 4f F3* 8 8 4281.32 0.006 0.001 0.007 0.001 ii V53b 3d 4P 4f D2* 6 6 4282.96 0.009 0.001 0.009 0.002 ii V67c 3d 4D 4f F2* 4 6 4283.73 0.009 0.002 0.007 0.002 ii V67c 3d 4D 4f F2* 4 4 4285.69 0.017 0.002 0.013 0.003 ii V78b 3d 2F 4f F3* 6 8 4287.39 0.074 0.007 0.081 [Fe ii] F7

Table 6: continued

Observed and dereddened emission line fluxes from IC 4776 (Å) Ion Multiplet Lower term Upper term g g 4291.25 0.011 0.002 0.011 0.002 ii V55 3d 4P 4f G3* 6 8 4292.16 0.007 0.002 0.008 0.001 ii 4f 2F* 10g 2G 14 18 4292.21 * * ii V78c 3d 2F 4f F2* 6 6 4294.78 0.019 0.002 0.020 0.001 ii V53b 3d 4P 4f D2* 4 6 4294.92 * * ii V53b 3d 4P 4f D2* 4 4 4303.82 0.027 0.002 0.027 0.002 ii V53a 3d 4P 4f D3* 6 8 4317.14 0.050 0.008 0.046 0.008 ii V2 3s 4P 3p 4P* 2 4 4319.63 0.039 0.003 0.040 0.004 ii V2 3s 4P 3p 4P* 4 6 4340.47 44.832 0.524 47.501 i H5 2p+ 2P* 5d+ 2D 8 50 4345.55 0.065 0.010 0.069 0.001 ii V2 3s 4P 3p 4P* 4 2 4345.56 * * ii V2 3s 4P 3p 4P* 4 2 4349.43 0.085 0.004 0.090 0.005 ii V2 3s 4P 3p 4P* 6 6 4351.81 0.097 0.013 0.078 0.014 [Fe ii] 4363.21 6.546 0.312 6.736 0.340 [O iii] F2 2p2 1D 2p2 1S 5 1 4366.89 0.052 0.005 0.056 0.005 iii V2 3s 4P 3p 4P* 6 4 4387.93 0.573 0.011 0.607 0.013 He i V51 2p 1P* 5d 1D 3 5 4409.30 0.013 0.001 0.013 0.001 Ne ii V55e 3d 4F 4f 2[5]* 8 10 4414.90 0.048 0.012 0.051 ii V5 3s 2P 3p 2D* 4 6 4416.97 0.026 0.014 0.049 0.015 ii V5 3s 2P 3p 2D* 2 4 4428.52 0.018 0.002 0.019 0.001 Ne ii V61b 3d 2D 4f 2[3]* 6 8 4428.64 * * Ne ii V60c 3d 2F 4f 1[3]* 6 8 4430.94 0.006 0.002 0.008 0.002 Ne ii V61a 3d 2D 4f 2[4]* 6 8 4432.74 0.006 0.002 0.006 0.001 ii V55b 3d 3P* 4f 2[3] 5 7 4432.75 * * ii V55b 3d 3P* 4f 2[3] 5 7 4437.55 0.072 0.004 0.073 0.004 He i V50 2p 1P* 5s 1S 3 1 4452.37 0.024 0.005 0.029 0.005 ii V5 3s 2P 3p 2D* 4 4 4471.50 4.723 0.070 4.888 0.083 He i V14 2p 3P* 4d 3D 9 15 4481.21 0.029 0.006 0.030 0.007 Mg ii V4 3d 2D 4f 2F* 10 14 4487.72 0.009 0.002 0.008 0.002 ii V104 3d’ 2P 4f’ D2* 2 4 4488.75 0.011 0.002 0.008 0.002 [Fe ii] F6 4s 6D 4s 4F 6 6 4491.07 0.011 0.001 0.011 0.001 ii 4f 2F* 9g 2G 14 18 4491.23 * * ii V86a 3d 2P 4f D3* 4 6 4510.91 0.009 0.002 0.008 0.002 iii V3 3s’ 4P* 3p’ 4D 2 4 4514.86 0.010 0.001 0.010 0.001 iii V3 3s’ 4P* 3p’ 4D 6 8 4518.15 0.003 0.001 0.005 0.001 iii V3 3s’ 4P* 3p’ 4D 2 2 4552.53 0.009 0.002 0.013 0.003 ii V58a 3d 1F* 4f 2[4] 7 9 4562.60 0.011 0.001 0.011 0.001 Mg i] 3s2 1S 3s3p 3P* 1 5 4571.10 0.176 0.014 0.168 Mg i] 3s2 1S 3s3p 3P* 1 3 4590.97 0.045 0.002 0.049 0.002 ii V15 3s’ 2D 3p’ 2F* 6 8 4596.18 0.037 0.004 0.036 0.004 ii V15 3s’ 2D 3p’ 2F* 4 6 4601.48 0.010 0.001 0.010 0.001 ii V5 3s 3P* 3p 3P 3 5 4602.13 0.010 0.001 0.009 0.001 ii V92b 3d 2D 4f F3* 4 6 4607.03 0.069 0.004 0.071 0.001 [Fe iii] F3 3d6 5D 3d6 3F2 9 7 4607.16 * * ii V5 3s 3P* 3p 3P 1 3 4609.44 0.018 0.003 0.022 0.003 ii V92a 3d 2D 4f F4* 6 8 4610.20 0.064 0.002 0.067 0.003 ii V92c 3d 2D 4f F2* 4 6 4613.68 0.006 0.001 0.006 0.001 ii V92b 3d 2D 4f F3* 6 8 4613.87 * * ii V5 3s 3P* 3p 3P 3 3 4621.25 0.012 0.003 0.012 0.001 ii V92 3d 2D 4f 2[2]* 6 6 4621.39 * * ii V5 3s 3P* 3p 3P 3 1 4630.54 0.020 0.001 0.023 0.002 ii V5 3s 3P* 3p 3P 5 5 4634.14 0.039 0.001 0.039 0.002 iii V2 3p 2P* 3d 2D 2 4 4638.86 0.066 0.003 0.072 0.003 ii V1 3s 4P 3p 4D* 2 4 4640.64 0.071 0.004 0.075 0.004 iii V2 3p 2P* 3d 2D 4 6 4641.81 0.157 0.004 0.161 0.001 ii V1 3s 4P 3p 4D* 4 6 4641.84 * * iii V2 3p 2P* 3d 2D 4 4 4647.42 0.010 0.002 0.009 0.002 iii V1 3s 3S 3p 3P* 3 5 4649.13 0.253 0.008 0.261 0.008 ii V1 3s 4P 3p 4D* 6 8 4650.84 0.055 0.006 0.064 0.006 ii V1 3s 4P 3p 4D* 2 2 4658.10 0.870 0.069 0.964 [Fe iii] F3 3d6 5D 3d6 3F2 9 9 4661.63 0.099 0.005 0.099 0.006 ii V1 3s 4P 3p 4D* 4 4 4669.27 0.004 0.001 0.004 0.001 ii V89b 3d 2D 4f D2* 4 6 4673.73 0.013 0.001 0.013 0.001 ii V1 3s 4P 3p 4D* 4 2 4676.24 0.056 0.002 0.057 0.002 ii V1 3s 4P 3p 4D* 6 6

Table 7: continued

Observed and dereddened emission line fluxes from IC 4776 (Å) Ion Multiplet Lower term Upper term g g 4685.68 0.018 0.006 0.020 0.006 He ii 3.4 3d+ 2D 4f+ 2F* 18 32 4696.35 0.008 0.001 0.007 0.001 ii V1 3s 4P 3p 4D* 6 4 4699.22 0.012 0.002 0.015 0.002 ii V25 3p 2D* 3d 2F 4 6 4701.62 0.359 0.022 0.335 0.022 [Fe iii] F3 3d6 5D 3d6 3F2 7 7 4711.37 0.158 0.005 0.158 0.005 [Ar iv] F1 3p3 4S* 3p3 2D* 4 6 4713.17 0.744 0.014 0.760 0.015 He i V12 2p 3P* 4s 3S 9 3 4733.91 0.152 0.007 0.152 0.008 [Fe iii] F3 3d6 5D 3d6 3F2 5 5 4740.17 0.485 0.022 0.514 0.023 [Ar iv] F1 3p3 4S* 3p3 2D* 4 4 4754.69 0.179 0.013 0.181 0.001 [Fe III] F 3 4754.70 * * [Fe III] F 3 4769.40 0.125 0.008 0.126 0.008 [Fe III] F 3 4788.13 0.008 0.001 0.009 0.001 ii V20 3p 3D 3d 3D* 5 5 4802.23 0.003 0.002 0.005 0.002 ii 4f 2F* 8g 2G 14 18 4803.29 0.011 0.002 0.009 0.002 ii V20 3p 3D 3d 3D* 7 7 4814.53 0.044 0.002 0.043 0.002 [Fe ii] F20 4815.55 0.007 0.002 0.006 0.002 ii V9 4s 4P 4p 4S* 6 4 4861.33 103.992 3.705 100.000 4.000 i H4 2p+ 2P* 4d+ 2D 8 32 4881.11 0.581 0.050 0.512 [Fe iii] F2 3d6 5D 3d6 3H 9 9 4906.83 0.028 0.005 0.029 0.005 [Fe iv] V28 3p 4S* 3d 4P 4 4 4921.93 1.299 0.042 1.262 0.042 He i V48 2p 1P* 4d 1D 3 5 4958.91 312.355 9.494 310.000 9.000 [O iii] F1 2p2 3P 2p2 1D 3 5 5006.84 952.817 27.604 938.000 27.000 [O iii] F1 2p2 3P 2p2 1D 5 5 5015.68 2.513 0.063 2.492 0.063 He i V4 2s 1S 3p 1P* * 1 5047.74 0.762 0.114 0.610 He i V47 2p 1P* 4s 1S 3 1 5191.82 0.088 0.003 0.084 0.003 [Ar iii] F3 2p4 1D 2p4 1S 5 1 5197.90 0.130 0.008 0.122 0.008 5200.26 0.082 0.005 0.074 0.005 [N i] F1 2p3 4S* 2p3 2D* 4 6 5270.40 0.531 0.052 0.518 [Fe iii] F1 3d6 5D 3d6 3P2 7 5 5342.38 0.010 0.002 0.010 0.002 ii 4f 2F* 7g 2G 14 18 5345.90 0.012 0.002 0.008 0.002 [Kr iv] F1 5411.52 0.055 0.005 0.048 He ii 4.7 4f+ 2F* 7g+ 2G 32 98 5453.83 0.006 0.002 0.007 0.002 ii V6 4s 4P 4p 4D* 6 8 5517.66 0.250 0.022 0.229 [Cl iii] F1 2p3 4S* 2p3 2D* 4 6 5537.60 0.602 0.031 0.571 0.029 [Cl iii] F1 2p3 4S* 2p3 2D* 4 4 5577.34 0.206 0.041 0.166 [O i] F3 2p4 1D 2p4 1S 5 1 5666.63 0.024 0.002 0.022 0.002 ii V3 3s 3P* 3p 3D 3 5 5676.02 0.012 0.002 0.010 0.002 ii V3 3s 3P* 3p 3D 1 3 5679.56 0.040 0.002 0.034 0.002 ii V3 3s 3P* 3p 3D 5 7 5686.21 0.007 0.002 0.005 0.001 ii V3 3s 3P* 3p 3D 3 3 5710.77 0.008 0.002 0.007 0.002 ii V3 3s 3P* 3p 3D 5 5 5754.60 1.045 0.089 0.909 [N ii] F3 2p2 1D 2p2 1S 5 1 5868.00 0.012 0.005 0.014 0.005 [Kr iv] 4p3 4S 3d3 2G 4 4 5875.66 16.512 0.382 15.170 0.427 He i V11 2p 3P* 3d 3D 9 15 5927.81 0.005 0.001 0.003 0.001 ii V28 3p 3P 3d 3D* 1 3 5931.78 0.018 0.004 0.015 0.003 ii V28 3p 3P 3d 3D* 3 5 5941.65 0.011 0.002 0.009 0.002 ii V28 3p 3P 3d 3D* 5 7 5978.97 0.058 0.003 0.050 0.003 iii V4 6101.83 0.031 0.002 0.025 0.002 [K iv] F1 3p4 3P 3d4 1D 5 5 6151.43 0.013 0.001 0.009 0.001 ii V16.04 4d 2D 6f 2F* 10 14 6233.80 0.019 0.003 0.011 0.003 6300.34 4.657 0.390 4.304 [O i] F1 2p4 3P 2p4 1D 5 5 6312.10 2.438 0.147 2.249 [S iii] F3 2p2 1D 2p2 1S 5 1 6347.10 0.101 0.005 0.084 0.005 Si ii V2 4s 2S 4p 2P* 2 4 6363.78 1.694 0.114 1.431 0.103 [O i] F1 2p4 3P 2p4 1D 3 5 6371.38 0.069 0.004 0.066 0.004 iii V2 4s 2S 4p 2P* 2 2 6461.95 0.013 0.003 0.018 0.002 ii 4f 2F* 6g 2G 14 18 6548.10 11.700 0.849 9.715 [N ii] F1 2p2 3P 2p2 1D 3 5 6560.10 0.140 0.042 0.115 He ii 4.6 4f+ 2F* 6g+ 2G 32 * 6562.77 345.244 6.736 288.000 5.000 i H3 2p+ 2P* 3d+ 2D 8 18 6583.50 39.739 3.370 31.100 [N ii] F1 2p2 3P 2p2 1D 5 5 6678.16 4.886 0.143 4.252 0.164 He i V46 2p 1P* 3d 1D 3 5 6716.44 1.734 0.074 1.379 [S ii] F2 2p3 4S* 2p3 2D* 4 6 6730.82 3.537 0.073 2.850 0.098 [S ii] F2 2p3 4S* 2p3 2D* 4 4 6795.00 0.006 0.001 0.005 0.001 [K iv] F1 3p4 3P 3p4 1D 3 5

Table 8: continued

Observed and dereddened emission line fluxes from IC 4776 (Å) Ion Multiplet Lower term Upper term g g 7065.25 11.038 0.254 8.808 0.340 He i V10 2p 3P* 3s 3S 9 3 7135.80 17.979 0.671 14.576 0.716 [Ar iii] F1 3p4 3P 3p4 1D 5 5 7160.56 0.033 0.001 0.026 0.001 He i 3s 3S 10p 3P* 3 9 7231.32 0.029 0.003 0.026 0.003 ii V3 3p 2P* 3d 2D 2 4 7236.19 0.061 0.004 0.049 0.002 ii V3 3p 2P* 3d 2D 4 6 7236.42 * * ii V3 3p 2P* 3d 2D 4 6 7237.17 0.024 0.007 0.019 0.001 ii V3 3p 2P* 3d 2D 4 4 7237.26 * * [Ar iv] F2 3p3 2D* 3p3 2P* 6 4 7254.38 0.136 0.009 0.105 i V20 3p 3P 5s 3S* 3 3 7262.76 0.014 0.003 0.011 0.002 [Ar iv] F2 3p3 2D* 3p3 2P* 4 2 7281.35 0.936 0.023 0.784 0.032 He i V45 2p 1P* 3s 1S 3 1 7298.04 0.041 0.003 0.032 0.003 He i 3s 3S 9p 3P* 3 9 7318.92 1.511 0.533 1.766 [O ii] F2 2p3 2D* 2p3 2P* 6 2 7319.99 8.096 0.592 6.517 [O ii] F2 2p3 2D* 2p3 2P* 6 4 7329.67 4.017 0.494 3.536 [O ii] F2 2p3 2D* 2p3 2P* 4 2 7330.73 3.958 0.566 3.202 [O ii] F2 2p3 2D* 2p3 2P* 4 4 7452.50 0.069 0.003 0.055 0.003 [Fe ii] F14 7499.84 0.062 0.003 0.046 0.003 He i 3s 3S 8p 3P* 3 9 7530.54 0.108 0.004 0.086 0.003 [Cl iv] 3p2 3P 3p2 1D 3 5 7530.83 * * [Cl iv] F1 3p2 3P 3p2 1D 3 5 7751.06 4.441 0.166 3.490 [Ar iii] 3p4 3P 3p4 1D 3 5 7751.12 * * [Ar iii] F1 3p4 3P 3p4 1D 3 5 7751.43 * * [Ar iii] F1 3p4 3P 3p4 1D 3 5 7816.16 0.092 0.003 0.072 0.003 He i V69 3s 3S 7p 3P* 3 9 8014.00 0.031 0.006 0.020 0.005 iii 3p’ 2D 3d’ 2D* 6 4 8045.63 0.244 0.006 0.190 0.009 [Cl iv] 3p2 3P 3p2 1D 5 5 8083.88 0.009 0.001 0.008 0.001 Fe ii z4D* d4P 4 6 8101.31 0.013 0.001 0.010 0.001 Fe iii b2D 4s3D 4P 4p5P* 5 7 8101.56 * * Ti iii 3d5f 1F* 6g 24 7 7 8116.60 0.011 0.003 0.010 0.002 He i 3.16 3p 3P* 16d 3D 9 8185.51 0.017 0.003 0.012 0.002 Ne i 4s 22* 7p 21 3 3 8204.42 0.013 0.005 0.011 0.004 Fe ii 4P sp4P* 3F 4d4P 2 4 8215.90 0.029 0.009 0.026 0.007 ii ? ? 8249.97 0.034 0.015 0.036 0.012 i P40 3d+ 2D 40f+ 2F* 18 8255.02 0.079 0.014 0.042 0.011 i P38 3d+ 2D 38f+ 2F* 18 8260.93 0.075 0.005 0.060 0.005 i P36 3d+ 2D 36f+ 2F* 18 8264.28 0.113 0.004 0.082 0.005 i P35 3d+ 2D 35f+ 2F* 18 8267.94 0.098 0.003 0.074 0.004 i P34 3d+ 2D 34f+ 2F* 18 8271.93 0.100 0.002 0.076 0.004 i P33 3d+ 2D 33f+ 2F* 18 8276.31 0.117 0.006 0.091 0.006 i P32 3d+ 2D 32f+ 2F* 18 8281.12 0.137 0.012 0.099 i P31 3d+ 2D 31f+ 2F* 18 8286.43 0.129 0.010 0.101 i P30 3d+ 2D 30f+ 2F* 18 8292.31 0.159 0.004 0.120 0.006 i P29 3d+ 2D 29f+ 2F* 18 8298.83 0.176 0.020 0.138 i P28 3d+ 2D 28f+ 2F* 18 8306.11 0.176 0.003 0.136 0.006 i P27 3d+ 2D 27f+ 2F* 18 8314.26 0.196 0.002 0.151 0.006 i P26 3d+ 2D 26f+ 2F* 18 8323.42 0.213 0.003 0.163 0.007 i P25 3d+ 2D 25f+ 2F* 18 8333.78 0.237 0.003 0.183 0.008 i P24 3d+ 2D 24f+ 2F* 18 8345.47 0.223 0.066 0.211 i P23 3d+ 2D 23f+ 2F* 18 8359.00 0.323 0.004 0.243 0.011 i P22 3d+ 2D 22f+ 2F* 18 8374.48 0.341 0.004 0.258 0.011 i P21 3d+ 2D 21f+ 2F* 18 8392.40 0.390 0.005 0.294 i P20 3d+ 2D 20f+ 2F* 18 8100.00 0.069 0.001 0.055 0.002 Paschen Jump- 8400.00 0.032 0.003 0.024 0.002 Paschen Jump+ 8413.32 0.436 0.018 0.342 0.020 i P19 3d+ 2D 19f+ 2F* 18 8421.99 0.020 0.002 0.014 0.001 He i 3.18 3d 3D 18p 3P* 15 8437.95 0.508 0.007 0.385 0.017 i P18 3d+ 2D 18f+ 2F* 18 8444.69 0.060 0.010 0.048 0.008 He i 3p 3P* 11d 3D 9 15 8451.20 0.056 0.014 0.058 0.011 He i 3.17 3d 3D 17p 3P* 15 8467.25 0.633 0.017 0.440 i P17 3d+ 2D 17f+ 2F* 18 8480.85 0.032 0.003 0.028 0.003 [Cl iii] 3p3 2D* 3p3 2P* 6 4 8486.31 0.024 0.001 0.018 0.001 He i 3.16 3d 3D 16p 3P* 15 8502.48 0.501 0.018 0.390 i P16 3d+ 2D 16f+ 2F* 18 8665.02 0.582 0.023 0.429 i P13 3d+ 2D 13f+ 2F* 18

Table 9: continued

Observed and dereddened emission line fluxes from IC 4776 (Å) Ion Multiplet Lower term Upper term g g 8680.28 0.029 0.002 0.022 0.001 i V1 3s 4P 3p 4D* 6 8 8680.53 * * Ca iii 5d 24* 5f 23 9 7 8686.15 0.025 0.008 0.018 0.006 i V1 3s 4P 3p 4D* 2 4 8703.87 0.013 0.005 0.012 0.004 [Ni ii] (3F)4s 2F (1D)4s 2D 8 6 8728.90 0.021 0.005 0.012 0.004 [Fe iii] 3P4 3D 5 7 8733.43 0.056 0.002 0.043 He i 3.12 3d 3D 12f 3F* 15 8739.97 0.003 0.001 0.002 0.001 He i 3d 3D 12p 3P* 5 3 8750.47 1.393 0.037 1.021 0.053 i P12 3d+ 2D 12f+ 2F* 18 8776.97 0.135 0.041 0.122 0.031 He i 3.9 3p 3P* 9d 3D 1 3 8809.83 0.003 0.002 0.005 0.001 Fe ii b4G x4G* 12 12 8816.64 0.008 0.001 0.006 He i 3p 1P* 12d 1D 3 5 8822.18 0.002 0.001 0.002 0.001 ii 4d 4F 5f G3* 6 8 8845.37 0.078 0.002 0.056 0.003 He i 3.11 3d 3D 11f 3F* 15 8862.78 1.758 0.037 1.304 0.065 i P11 3d+ 2D 11f+ 2F* 18 8899.31 0.005 0.001 0.003 0.001 i 3d 1D* 8p 1D 5 5 8914.77 0.017 0.001 0.012 0.001 He i 3.7 3s 1S 7p 1P* 1 3 8957.25 0.158 0.031 0.141 Si iv V3.01 5s 2S 5p 2P* 2 4 8997.02 0.106 0.003 0.074 0.004 He i 3.10 3d 3D 10f 3F* 15 9014.91 2.100 0.026 1.560 0.074 i P10 3d+ 2D 10f+ 2F* 18 9068.60 36.694 1.484 28.800 1.700 [S iii] 3p2 3P 3p2 1D 3 5 9084.60 0.006 0.002 0.008 0.002 Ne ii 3d 4F 4p 4D* 6 4 9095.46 0.010 0.002 0.009 0.002 i 3d 3D* 9p 3D 5 7 9123.60 0.035 0.007 0.027 0.005 [Cl ii] 3p4 3P 3p4 1D 3 5 9174.49 0.007 0.002 0.006 0.002 He i 3.8 3p 3P* 8s 3S 9 3 9210.33 0.131 0.004 0.100 He i V83 3d 3D 9f 3F* 15 21 9229.01 3.056 0.089 2.269 i P9 3d+ 2D 9f+ 2F* 18 9323.06 0.188 0.053 0.139 Fe ii e4D 5D 5p4P* 4 6 9402.17 0.051 0.005 0.037 0.004 Ne ii 4f 24* 5g 15 10 12 9463.58 0.259 0.052 0.178 He i V67 3s 3S 5p 3P* 3 9 9475.03 0.266 0.083 0.207 Ni ii 4f 33* 3F 6d2G 8 10 9516.63 0.100 0.018 0.081 He i V76 3p 3P* 7d 3D 9 15 9530.60 92.480 3.012 67.600 [S iii] 3p2 3P 3p2 1D 5 5 9545.97 3.478 0.610 2.674 i P8 3d+ 2D 8f+ 2F* 18 9603.44 0.025 0.006 0.014 0.004 He i V71 3s 1S 6p 1P* 1 3 9620.50 0.059 0.009 0.038 0.007 iii 3H* 3G 33 27 9702.71 0.039 0.014 0.030 0.010 He i V75 3p 3P* 7s 3S 9 3 9720.13 0.145 0.029 0.105 Ne ii 4p 2S* 4d 2P 2 2 9790.52 0.256 0.065 0.228 Fe ii 4f 46* 5g 47 14 16 9799.26 0.094 0.013 0.072 Ne ii 2P* 2P 4 4 9824.13 0.013 0.003 0.009 0.002 [C i] 2p2 3P 2p2 1D 3 5 9868.21 0.056 0.007 0.034 0.005 ii V65.09 4f 13 5g 14* 5 7 9891.09 0.025 0.009 0.029 0.007 ii V65.11b 4f 14 5g 15* 18 27 9903.39 0.078 0.026 0.056 0.003 ii 4d 2D 5f F3* 6 6 9903.45 * * ii V17.02 4f 2F* 5g 2G 22 26 9903.46 * * ii V17.02 4f 2F* 5g 2G 22 26 9911.46 0.042