ON THE PROGENITORS OF GALACTIC NOVAE

A classical nova (CN) outburst occurs in an interacting binary system comprising a white dwarf (WD, the primary) and typically a late-type main-sequence (MS) star (the secondary) that fills its Roche lobe (Crawford & Kraft 1956). Matter is transferred, usually via an accretion disk around the WD, from the secondary to the WD surface, becoming degenerate in the process. Once extensive CNO cycle nuclear burning commences under these degenerate conditions, a thermonuclear runaway occurs ejecting ∼10⁻⁵–10⁻⁴ M_☉ of matter from the system with velocities from hundreds to thousands of km s⁻¹ (see, e.g., Bode & Evans 2008, and references therein). These systems generally have orbital periods of a few hours (Warner 2008) and recurrence times of a few × 10³–10⁶ years are often quoted (Truran & Livio 1986). CNe are traditionally sub-divided using solely their eruptive properties, typically speed class (Gaposchkin 1957), or spectral class (Williams 1992). More recently, novae have also been classified as "disk" or "bulge" novae, on the basis of their parent stellar population and location within the host galaxy (Duerbeck 1990; Della Valle et al. 1992; Della Valle & Livio 1998).

Closely related to these systems are the recurrent novae (RNe), which, unlike CNe, are observed to recur on timescales of order tens to a hundred years. However, it should be noted that the upper limit to this range is likely to be an observational effect based on the paucity of historical records. Such a short recurrence timescale is usually attributed to a high-mass WD, probably close to the Chandrasekhar limit, together with a high accretion rate (Starrfield et al. 1985; Yaron et al. 2005). The orbital periods of RNe are less homogeneous than those of CNe, with periods ranging from a few hours to several hundreds of days (Anupama 2008).

RN systems have been sub-divided into three general sub-classes (Anupama 2008) via a combination of the properties of the eruption and via the properties of the progenitor system while at quiescence:

RS Oph/T CrB systems (hereafter the RS Oph-class) are observed to contain red giant secondaries and hence have much longer orbital periods (∼a year), and typically smaller outburst amplitudes than CNe. Their outbursts exhibit rapid declines from maximum, with large ejection velocities (⩾4000 km s⁻¹). These systems show evidence of an interaction between the ejecta and material from the pre-existing red giant wind. The ejected mass from these systems is typically two orders of magnitude less than that observed from CN systems (see papers within Evans et al. 2008).

U Sco systems are observed to contain evolved MS or sub-giant secondaries, with orbital periods closer to those of CNe (hours up to of order a day). They again exhibit rapid declines, being among the fastest declining novae observed, with very high ejection velocities (up to 10, 000 km s⁻¹). Their post-outburst spectra resemble those of the He/N sub-class of CNe (Williams 1992). These systems eject a similar mass of material as RS Oph systems.

The T Pyx systems (also CI Aql and IM Nor) are much more akin to CNe. They exhibit short orbital periods and spectroscopically resemble Fe ii CNe (Williams 1992). Their optical decline rate classifies them as moderately fast or slow novae. The ejected mass in these systems is consistent with the range observed in CNe (M_ej ∼ 10⁻⁵ M_☉). These systems are generally only distinguishable from CNe due to their shorter recurrence times, not by the properties of the progenitor system or the outburst.

The short inter-outburst time observed in RNe is likely due to some combination of a higher mass WD and an accretion rate greater than the typical for CN systems. Indeed, both RS Oph and U Sco (among others from these classes) appear to have WDs close to the Chandrasekhar limit. A number of authors (see, e.g., Hachisu et al. 2007; Osborne et al. 2011; Starrfield et al. 2011) have indicated that the WD mass may be increasing over time and these systems have been proposed as a Type Ia supernova (SN Ia) progenitor candidate (see, e.g., Kotak 2008). However, Mason (2011) reported that U Sco could contain an O-Ne WD, rather than a C-O WD, and hence may not evolve in to an SN Ia as is widely predicted for this system.

Any CN that has been observed to erupt more than once is reclassified as an RN. This approach has given us 10 Galactic RNe among ∼400 known Galactic CNe from an underlying rate of 34⁺¹⁵₋₁₂ year⁻¹ (Darnley et al. 2006). Due to observational effects—which typically worsen the further back in time one looks—it is highly likely that a considerable proportion of these Galactic CNe are in fact (short timescale) recurrents for which only a single outburst has been observed. Either a previous outburst has been missed or the inter-outburst period is particularly long. So it would be seem advantageous to better develop a nova classification system that relies more on the fundamental properties of the systems than on observational selection effects.

Over the past few years observations of a number of Galactic "transients" have shown properties consistent with being RNe, although they had never been seen in outburst previously. For example, the nova V2487 Ophiuchi (Nova Oph 1998) was observed to have a rapid optical decline and plateau phase (Hachisu et al. 2002; Hernanz & Sala 2002), suggestive of an RN. Consultation of the Harvard College Observatory archival photographic collection revealed a previously unknown outburst on 1900 June 20 (Pagnotta et al. 2009). Additionally, the novae V1721 Aql (Hounsell et al. 2011), V2491 Cyg (Darnley et al. 2011), KT Eri (Jurdana-Šepić et al. 2012), and V2672 Oph (Munari et al. 2011) and also the extragalactic nova M31N 2007-12b (Bode et al. 2009) have been shown to harbor secondary stars akin to either the U Sco or RS Oph systems, albeit with only one recorded outburst.

In this paper, we will attempt to simplify the nomenclature somewhat by only classifying each system using the evolutionary state of the secondary; an MS star (MS-Nova), a sub-giant star (SG-Nova), or a red giant branch (RGB) star (RG-Nova). We introduce such classifications in an attempt to avoid the accretion rate/WD mass/recurrence time degeneracy. All the known U Sco-class RNe would be placed into the SG-Nova group, all the known RS Oph-class RNe into the RG-Nova group, the T Pyx-class RNe and Classical Novae will populate the MS-Nova group.

By virtue of the geometry of these systems, the above classifications would effectively be the same as sub-dividing novae by orbital period, if one assumes that the secondaries fill their Roche lobes. The MS-Novae would have orbital periods of order hours, the SG-Novae of order a day, and the RG-Novae of order a year. Indeed, Warner (1995) indicates that systems with orbital periods longer than 8 hr should contain evolved secondaries.

In this paper, we present a simple method which will allow differentiation of RG-Novae and SG-Novae from the MS-Novae population based solely on their quiescent (inter-outburst) broadband optical and near-IR (NIR) properties. The format of the remainder of this paper is as follows. In Section 2 we briefly describe our method and present the data used to illustrate this technique; in Section 3 we present our results, in Section 4 we discuss our findings and present a number of predictions and finally, in Section 5 we summarize our conclusions.

The broadband optical emission from all quiescent nova systems is a superposition of the emission from the three main components of the system: the WD, the accretion disk, and the secondary. In all cases we would expect the WD's contribution in the optical to be negligible. The contribution of the accretion disk to the emission is a combination of a number of factors including accretion rate, disk size, system inclination, and wavelength. Whereas the contribution from the secondary is much more straightforward and simply depends upon the type (mass, age, and metallicity) of the star and wavelength.

Both the SG-Novae and RG-Novae have much smaller optical outburst amplitudes than MS-Novae, although the energetics of the outbursts of all three systems are comparable. In the RG-Nova systems this can be readily explained by the presence of a—highly luminous—red giant secondary, possibly with some additional contribution from a disk, especially in bluer optical filters. For the SG-Nova systems the cause is not as clear-cut but must be some combination of a—luminous—sub-giant secondary and an accretion disk, possibly experience a high-mass accretion rate.

The broadband NIR emission from quiescent RG-Nova and SG-Nova systems should be dominated by the emission from the cool secondary stars, with a small contribution from the accretion disk. At these wavelengths, the accretion disk emission may still be relatively significant for MS-Nova systems due to the low luminosity of the dwarf secondary star.

As such, we propose that the position of a quiescent nova system on a color–magnitude diagram can be broadly predicted. Each system would be expected to lie at such a position that it appears more luminous and bluer (hotter) than the secondary in the system, with the magnitude of the luminosity and temperature increase being a function of both the accretion rate, system inclination and—of course—wavelength. Conversely, by plotting quiescent nova systems on a color–magnitude diagram one can identify the type of secondary and hence predict whether a given system may be an RG-Nova, SG-Nova, or MS-Nova system.

In Table 1, we present quiescent broadband optical and NIR photometric data for a sample of Galactic novae (and a number of extragalactic examples). These data have been collated from a wide range of sources, with all the NIR data taken from the Two Micron All Sky Survey (2MASS; Skrutskie et al. 2006) either directly from the catalog or indirectly through additional sources. In Table 2, we provide a summary of the best determination of the line-of-sight extinction and distance to each system as well as the system inclination (if known). This catalog of Galactic novae was compiled by virtue of the existence of quiescent photometry in at least two optical bands, and reasonable determinations of both distance and extinction.

Table 1. Optical and NIR Photometry of Quiescent Nova Systems

Nova	System	Recurrent	Secondary	B	V	R	I	J	H	Ks
Name	Typea		Spectral	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)
			Classb					2MASS photometryc
CI Aql	SG-Nova (1)	$\checkmark$	K-M IV	17.10 (2)	16.10 (3)	15.34 (2,3)	...	13.67 ± 0.04	>13.33 ± 0.08	>12.7
V356 Aql	MS-Nova			18.50 (3)	18.30 (3)	17.80 (3)	...	...	...	...
V528 Aql	MS-Nova			19.25 (3)	18.48 (3)	17.92 (3)	...	...	...	...
V603 Aql	MS-Nova			11.63 (4)	11.67 (4)	...	...	11.70 ± 0.03	11.51 ± 0.03	11.35 ± 0.03
V1229 Aql	MS-Nova			19.73 (3)	18.59 (3)	17.74 (3)	...	16.3 ± 0.1	15.7 ± 0.1	>15.5
V1721 Aql	SG-Nova (5)			...	...	...	...	16.6 ± 0.2 (5)	15.5 ± 0.1 (4)	14.7 ± 0.1 (5)
T Aur	MS-Nova			15.20 (4)	14.92 (4)	...	...	14.01 ± 0.03	13.74 ± 0.03	13.58 ± 0.03
IV Cep	MS-Nova			17.02 (3)	16.40 (3)	...	...	14.96 ± 0.04	14.63 ± 0.07	14.30 ± 0.08
V394 CrA	SG-Nova	$\checkmark$	K	19.20 (2)	18.40 (2)	...	...	...	...	...
T CrB	RG-Nova	$\checkmark$	M3 III	11.60 (2)	10.10 (2)	...	...	6.00 ± 0.02	5.15 ± 0.04	4.81 ± 0.02
V407 Cyg	Sym-Nova			...	...	...	...	5.70 ± 0.02	4.36 ± 0.02	3.2 ± 0.3
V476 Cyg	MS-Nova			17.24 (4)	17.09 (4)	17.24 (3,4)	...	16.0 ± 0.1	15.6 ± 0.1	15.6 ± 0.2
V1500 Cyg	MS-Nova			17.44 (4)	17.06 (4)	15.15 (3)	...	...	...	...
V1974 Cyg	MS-Nova			16.88 (6)	16.63 (6)	...	...	15.67 ± 0.07	15.3 ± 0.1	15.2 ± 0.2
V2491 Cyg	SG-Nova (7)			18.34 (8)	17.88 (8)	17.49 (7)	17.14 (8)	16.7 ± 0.2 (7)	16.6 ± 0.3 (7)	16.7 ± 0.6 (7)
HR Del	MS-Nova			12.15 (4)	12.11 (4)	...	...	12.32 ± 0.02	12.28 ± 0.02	12.22 ± 0.03
KT Eri	RG-Nova (9)			15.2 ± 0.3 (9)	...	15.3 ± 0.3 (9)	14.7 ± 0.3 (9)	14.62 ± 0.03	14.16 ± 0.05	14.09 ± 0.07
DN Gem	MS-Nova			15.76 (4)	15.56 (4)	15.54 (3,4)	...	15.43 ± 0.05	15.24 ± 0.09	15.3 ± 0.2
DQ Her	MS-Nova (IP)			14.59 (4)	14.45 (4)	...	...	13.60 ± 0.03	13.28 ± 0.04	13.09 ± 0.04
V533 Her	MS-Nova			15.78 (4)	15.60 (4)	...	...	14.71 ± 0.03	14.65 ± 0.05	14.6 ± 0.1
CP Lac	MS-Nova			16.05 (3,4)	15.76 (4)	15.57 (3,4)	...	15.03 ± 0.06	14.73 ± 0.07	14.8 ± 0.1
DK Lac	MS-Nova			16.83 (3)	17.75 (3)	16.59 (3)	...	16.4 ± 0.1	16.0 ± 0.2	>15.9
DI Lac	SG-Nova (1)			14.93 (3)	14.68 (3)	...	...	13.73 ± 0.04	13.54 ± 0.05	13.40 ± 0.04
IM Nor	MS-Nova	$\checkmark$		19.30 (2)	18.30 (2)	...	...	15.46 ± 0.09	>13.6	>13.4
RS Oph	RG-Nova	$\checkmark$	M0/2 III	12.30 ± 0.03 (10,11)	10.98 ± 0.03 (10,11)	9.50 ± 0.03 (10,11)	8.67 ± 0.02 (10,11)	7.64 ± 0.02	6.86 ± 0.04	6.50 ± 0.02
V849 Oph	MS-Nova			19.30 (3)	18.80 (3)	18.50 (3)	...	...	...	...
V2487 Oph	RG-Nova (1)	$\checkmark$		18.10 (2)	17.30 (2)	...	...	15.36 ± 0.08	14.9 ± 0.1	14.41 ± 0.09
GK Per	SG-Nova (IP)			13.81 (4)	13.02 (4)	...	...	10.86 ± 0.02	10.27 ± 0.02	10.06 ± 0.02
RR Pic	MS-Nova			12.18 (4)	12.21 (4)	...	...	12.46 ± 0.02	12.40 ± 0.02	12.25 ± 0.02
CP Pup	MS-Nova			15.17 (4)	14.97 (4)	...	...	14.34 ± 0.03	14.24 ± 0.04	14.02 ± 0.07
T Pyx	MS-Nova	$\checkmark$		15.60 (2)	15.60 (2)	...	...	15.15 ± 0.05	14.96 ± 0.09	14.67 ± 0.09
U Sco	SG-Nova	$\checkmark$	K2 IV	18.55 (12)	18.01 (12)	17.67 (12)	17.21 (12)	16.8 ± 0.1	16.4 ± 0.2	>15.2
V745 Sco	RG-Nova	$\checkmark$	M6 III	20.50 (2)	18.70 (2)	...	...	10.04 ± 0.03	8.85 ± 0.04	8.26 ± 0.04
EU Sct	RG-Nova (1)			20.02 (3)	17.37 (3)	18.80 (3)	...	12.26 ± 0.05	11.18 ± 0.04	10.80 ± 0.04
FH Ser	MS-Nova			16.18 (4)	15.06 (4)	14.79 (3,4)	...	15.25 ± 0.06	15.0 ± 0.1	14.7 ± 0.1
V3890 Sgr	RG-Nova	$\checkmark$	M5 III	16.40 (2)	15.50 (2)	...	...	9.83 ± 0.02	8.77 ± 0.03	8.26 ± 0.02
LV Vul	MS-Nova			16.56 (3)	15.84 (3)	15.24 (3)	...	...	...	...
NQ Vul	MS-Nova			18.27 (3)	17.22 (3)	16.32 (3)	...	15.03 ± 0.05	14.63 ± 0.06	14.26 ± 0.08
LMCN 2009-02a	SG-Nova (13)	$\checkmark$(13)		18.3 ± 0.4(13)	...	19.2 ± 0.6 (13)	...	...	...	...
M31N 2007-12b	RG-Nova (10)			...	24.61 ± 0.09 (10)	...	22.33 ± 0.04 (10)	...	...	...

Notes. ^aStandard classification for each system, unless otherwise stated. ^bSpectral classes of the secondary stars of known RNe from Anupama (2008). ^cUnless otherwise stated, all NIR photometry is taken directly from the 2MASS catalog (Skrutskie et al. 2006). References. (1) This paper; (2) Schaefer 2010; (3) Szkody 1994; (4) Bruch & Engel 1994; (5) Hounsell et al. 2011; (6) Shugarov et al. 2002; (7) Darnley et al. 2011; (8) Munari et al. 2011; (9) Jurdana-Šepić et al. 2012; (10) Bode et al. 2009; (11) Darnley et al. 2008; (12) Schaefer et al. 2010; (13) M. F. Bode et al. (2012, in preparation).

Download table as:  ASCIITypeset image

In Figure 1, we present a range of color–magnitude plots. The positions of all the novae listed in Table 1 are plotted with appropriate error bars (or error estimations). For comparison, we have also plotted the local Galactic stellar population, generated from the Hipparcos catalog (Perryman & ESA 1997). Only those stars that have both photometric and parallax errors <10% have been selected from the Hipparcos data. The BVI photometry for these Hipparcos stars has been taken directly from that catalog, while the R and NIR photometry has been generated by cross-correlating the Hipparcos catalog with the NOMAD1 data set (Zacharias et al. 2004) and the 2MASS catalog, respectively. Included in each figure are the stellar evolutionary tracks of a 1 M_☉ solar-like star and a 1.4 M_☉ solar-like star (Pietrinferni et al. 2004) from the zero-age MS up to the tip of the RGB. The majority of the nova secondary masses are expected to be <1 M_☉, and for stable mass transfer the mass of the secondary must be <M_WD. We have provided four alternative color–magnitude diagrams: a standard diagram V versus B − V; as well as J versus B − J to particularly focus on the luminosity difference between the disk and secondary; and a pair of NIR diagrams to focus specifically on the secondary.

Zoom In Zoom Out Reset image size

Taking each plot from Figure 1 in turn.

• 1.  
  V versus B − V. This plot was included as this is the "standard" color–magnitude diagram and the most commonly used filters, particularly for observations of novae. Although there does seem to be some separation by position of the three secondary sub-types of nova systems, it is not overly clear. Unless the secondary star is particularly bright, or the system is very close to edge-on, the B and V photometry is strongly influenced by the accretion disk, and as such gives little information regarding the nature of the secondary star. As can be seen, a number of RG-Novae (red points) are interdispersed among the SG-Novae (green points) probably due to a combination of low system inclination angles and high accretion rates. However, the MS-Novae do appear to be separated from the SG-/RG-Novae by the MS itself, with the MS-Novae all lying on the blue side.
• 2.  
  J versus B − J. This plot was included as it specifically samples both the secondary star (J) and the accretion disk (B) giving some indication of the relative luminosity of the accretion disk in each system, and hence the accretion rate. As in the V versus B − V plot, a number of points have been redistributed significantly "blueward" from the position of the secondary.
• 3.  
  J versus J − H. The first of two NIR only color–magnitude diagrams. Such plots are expected to focus strongly on the secondary star which is expected to dominate the emission at these wavelengths for all but the lowest luminosity secondaries. As can be seen from the plot in Figure 1 there is a clear separation between the RG-Novae and the MS-Novae, and possibly a small distinct region belonging to the SG-Novae.
• 4.  
  J versus J − K_S. As expected, this plot is very similar to the J − H plot. This plot is slightly let down by the lower sensitivity of the 2MASS K_S data set; however, the separation between the three sub-classes is just as well defined.

The plots created for Figure 1 are also reproduced in Figure 2, however, in this figure we only plot (and identify) a selection of specific novae. We will now briefly discuss each of these selected novae.

Zoom In Zoom Out Reset image size

3.1. CI Aquilæ

3.2. KT Eridani

3.3. DI Lacertæ

3.4. V2487 Ophiuchus

3.5. GK Persei

3.6. V3890 Sagittarii

3.7. EU Scutum

As can be seen in Figure 1, when quiescent novae are plotted on color–magnitude diagrams, the position of a given system is a strong function of both the evolutionary state of the secondary star and the luminosity of any accretion disk. When we focus only on NIR observations, any dependency on the accretion disk is minimized and in many cases is negligible. Those systems known to harbor red giant secondary stars are correlated strongly with the RGB. Those containing MS secondaries are clustered blueward of the MS (due to the effect of the accretion disk) yet are less luminous than the (current) position of the local Galactic sub-giant branch. While less well defined, the locus of the sub-giant secondary systems is typically more luminous than the sub-giant branch and significantly bluer than the RGB.

This work has allowed us to reach a number of conclusions and predictions. First, we can identify a number of systems structurally akin to either U Sco (SG-Novae) or RS Oph (RG-Novae) that are only known to have undergone a single outburst. These SG-Novae are V1721 Aql and V2491 Cyg; the RG-Novae are KT Eri, EU Sct, and M31N 2007-12b. However, the question of whether these are RNe with missed outbursts or part of a population of long inter-outburst timescale evolved systems is one we cannot address here as a short recurrence time also requires a high-mass WD. However, we would strongly urge searches of observational archives to uncover any missed outbursts (as was done for V2487 Oph). We predict that the secondary of the RN CI Aql is a sub-giant star, akin to the secondary in the U Sco system, and that the secondary of the recurrent V2487 Oph is a red giant similar to that in the KT Eri system. There is also some evidence that the secondary in DI Lac is also a sub-giant star.

While preparing the catalog of novae analyzed in this paper, it quickly became clear that reliable, multi-color, quiescent photometry is lacking for a large number of Galactic novae. Additionally, reliance on a single epoch of quiescent photometry is problematic due to the large range in quiescent luminosity exhibited by many systems.

While this technique has been shown to produce results generally consistent with the current understanding of most nova systems, although revealing several interesting systems, there are a number of limitations that are worth of discussion.

4.1. Inclination

In systems where the accretion disk has a significant contribution to the luminosity, the magnitude of this contribution is a strong function of the inclination of that disk (or of the system). Warner (1987) showed a correlation between the disk inclination and the apparent brightness of nova remnants and the amplitude of their outburst. The inclination correction to be applied to the magnitude of a quiescent CN system was initially derived for the dwarf nova U Geminorum by Paczynski & Schwarzenberg-Czerny (1980). They modeled the disk luminosity in U Gem by assuming a flat optically thick accretion disk and derived the following relation showing the change in magnitude of the disk as a function of inclination:

$$\begin{equation} \Delta M_V(i) = -2.5 \log \left(\cos i+\frac{3}{2}\cos ^{2}i\right),\; 0^{\circ }\le i<90^{\circ } \hbox{,} \end{equation} \tag{ 1 }$$

where i is the inclination of the system, defined such that i = 0° is a polar/face-on system, and i = 90° is edge-on. This relationship is independent of the disk luminosity and hence the accretion rate and is normalized to the mean disk luminosity (which occurs at i = 567). We adapt Equation (1) such that it may be normalized to any inclination:

$$\begin{equation} \Delta M_{v}(i,{\cal I}) = -2.5\log \left(\frac{2\cos {i} + 3\cos ^{2} i}{2\cos {{\cal I}} + 3\cos ^{2} {\cal I}} \right),\; 0^{\circ }\le i,{\cal I}<90^{\circ } \hbox{,} \end{equation} \tag{ 2 }$$

where ${\cal I}$ is the inclination at which one chooses as a standard. The Paczynski & Schwarzenberg-Czerny (1980) model is applicable for any system in which the accretion disk is the dominant luminosity source and is therefore only reasonable to apply to MS-Nova systems, the effect of the magnitude change would be expected to decrease significantly for SG-Novae and RG-Novae, i.e., with increasing secondary luminosity. To investigate the effect of inclination in each nova system we use Equation (2) to minimize the effect of the accretion disk (i.e., edge-on). However, as the accretion disks in novae are not flat and the secondary has some (non-zero) luminosity in V (original assumptions from Paczynski & Schwarzenberg-Czerny 1980) we adopt an inclination ${\cal I}=80^{\circ }$ to be indicative of the minimum V-band flux of an edge-on CN system.

Therefore, to estimate the absolute magnitude M_V of the secondary in a given system at quiescence, it follows that

$$\begin{equation} M_{V}(i,{\cal I}) = V - 5 \log d + 5 - A_V - \Delta M_V(i,{\cal I}) \hbox{,} \end{equation} \tag{ 3 }$$

where V is the apparent magnitude, d is distance to the system in (parsecs), and A_V is the extinction correction. However, such an approach is only likely to be valid when it can be assumed that the accretion disk dominates the luminosity and the accretion rate is low (hence the disk is approximately flat). Hence, in SG-Novae and particular RG-Novae systems, where the flux contribution of the secondary is significant, we would expect the effect of system inclination to be substantially reduced. Additionally, the luminosity ratio of the accretion disk to the secondary will be at its smallest for the K_S data. Hence, if we assume that the K_S flux of each quiescent nova system is due only to the secondary, we can produce a color–magnitude diagram showing the location of just the secondary star in each system (see Figure 3).

Zoom In Zoom Out Reset image size

As can be seen in Figure 3 this very simple approach to inclination has the expected effect of repositioning most MS-nova systems redward, toward the true position of the secondary star on the MS. However, in order to apply such a technique to all novae, more detailed models of the disks in these systems are required. Such attempts also suffer from the difficulty in obtaining, and hence lack of, accurate inclination measurements of most nova systems.

4.2. Distance

4.3. Extinction

4.4. Orbital Period

In Figure 4, we present a plot of the known orbital periods of quiescent nova systems against that system's quiescent K_S-band luminosity. The distribution of the data shown in this plot is broadly as expected; the lower the K_S-band luminosity of the system, the shorter the orbital period. That is, the systems with the more evolved secondaries must have larger orbital separations between the primary and the secondary. However, there are a number of interesting outliers to the general trend.

Zoom In Zoom Out Reset image size

V2487 Oph appears particularly overluminous for such a short (∼1 day) period, based on the other RG-Nova systems we would expect a significantly longer period. It is interesting to note that although Schaefer (2010) indicates an orbital period of ∼1 day for V2487 Oph, he also states "no significant periodic modulation was seen (in the light curve)." As such, we would strongly recommend additional observations in order to directly determine the orbital period of this system.

Although the orbital period of the RG-Nova KT Eri is similar to others in that group, its luminosity is significantly lower. As such, we expect the secondary in this system to be less-evolved and hence physically smaller than, for example, RS Oph. This secondary must be far from being Roche lobe filling as was also suggested by Jurdana-Šepić et al. (2012) based on the period-folded quiescent light curve.

Our study of the archival optical and NIR photometry of a sample of quiescent Galactic classical and RNe have led us to the following primary conclusions.

• 1.  
  We have proposed a new CN classification system, based solely on the evolutionary state of the secondary star: an MS star (MS-Nova), a sub-giant star (SG-Nova), or an RGB star (RG-Nova).
• 2.  
  The position of a quiescent novae on optical or NIR color–magnitude diagrams is a strong function of both the evolutionary state of the secondary star and the luminosity of any accretion disk. The more evolved the secondary star, the weaker the position dependency is on the accretion disk.
• 3.  
  A standard (V versus (B − V)) color–magnitude diagram is useful for differentiating between the MS-Novae (which all lie bluer than the MS) and the SG-/RG-Novae (which all lie on or are redder than the MS). Although this plot can be strongly sensitive to the accretion rate.
• 4.  
  Additionally NIR color–magnitude diagrams may be used to differentiate between all three sub-classes of secondaries.

We also make a number of predictions and observations about systems included in this paper.

• 1.  
  The RN CI Aql—often grouped with T Pyx—is in fact an SG-Nova, a system more akin to U Sco.
• 2.  
  The RN V2487 Oph is a system similar to KT Eri, containing a low-luminosity RGB star. This contradicts the short orbital period reported in Schaefer (2010) and we recommend follow-up observations to directly determine the orbital period. Additionally, we cannot rule out the possibility that the secondary in this system is a horizontal branch star.
• 3.  
  Based on the high quiescent luminosity and long orbital period, we predict that the DI Lac system may contain a sub-giant secondary star.
• 4.  
  In line with Weight et al. (1994) and Woudt & Warner (2003), we propose that EU Sct contains a luminous red giant secondary.

Follow-up (quiescent) spectroscopic observations of novae, such as those initially reported in Surina et al. (2012), can greatly aid in the classification of the secondary stars in these systems.

This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation. This research has made use of NASA's Astrophysics Data System. This research has made use of the VizieR catalog access tool, CDS, Strasbourg, France. V.A.R.M.R. was supported by an STFC PhD studentship and is currently supported by a South African SKA Fellowship. R.A.H. acknowledges PhD funding from STFC. The authors thank the referee, Massimo Della Valle, for his very constructive comments.