Searching for nova shells around cataclysmic variables
MMon. Not. R. Astron. Soc. , 000–000 (0000) Printed 13 September 2018 (MN L A TEX style file v2.2)
Searching for nova shells around cataclysmic variables
D. I. Sahman, (cid:63) V. S. Dhillon, C. Knigge, T. R. Marsh Department of Physics and Astronomy, University of Sheffield, Sheffield S3 7RH, UK School of Physics & Astronomy, University of Southampton, Southampton SO17 1BJ, UK Department of Physics, University of Warwick, Coventry CV4 7AL, UK
Accepted 2015 May 16. Received 2015 April 17
ABSTRACT
We present the results of a search for nova shells around 101 cataclysmic variables(CVs), using H α images taken with the 4.2-m William Herschel Telescope (WHT) andthe 2.5-m Isaac Newton Telescope Photometric H α Survey of the Northern GalacticPlane (IPHAS). Both telescopes are located on La Palma. We concentrated our WHTsearch on nova-like variables, whilst our IPHAS search covered all CVs in the IPHASfootprint. We found one shell out of the 24 nova-like variables we examined. Thenewly discovered shell is around V1315 Aql and has a radius of ∼ . (cid:48) , indicative ofa nova eruption approximately 120 years ago. This result is consistent with the ideathat the high mass-transfer rate exhibited by nova-like variables is due to enhancedirradiation of the secondary by the hot white dwarf following a recent nova eruption.The implications of our observations for the lifetime of the nova-like variable phaseare discussed.We also examined 4 asynchronous polars, but found no new shells around anyof them, so we are unable to confirm that a recent nova eruption is the cause of theasynchronicity in the white dwarf spin. We find tentative evidence of a faint shellaround the dwarf nova V1363 Cyg. In addition, we find evidence for a light echoaround the nova V2275 Cyg, which erupted in 2001, indicative of an earlier novaeruption ∼
300 years ago, making V2275 Cyg a possible recurrent nova.
Key words: stars: novae, cataclysmic variables.
Cataclysmic variables (CVs) are close binary systems inwhich a white dwarf (WD) accretes material from a sec-ondary star, via Roche-lobe overflow (see Warner 1995 fora review). CVs are classified observationally into 3 mainsub-types – the novae, the dwarf novae and the nova-likevariables. The novae are defined as systems in which only asingle nova eruption has been observed. Nova eruptions havetypical amplitudes of 10 magnitudes and are believed to bedue to the thermonuclear runaway of hydrogen-rich mate-rial accreted onto the surface of the white dwarf. The dwarfnovae (DNe) are defined as systems which undergo quasi-regular (on timescales of weeks-months) outbursts of muchsmaller amplitude (typically 6 magnitudes). Dwarf nova out-bursts are believed to be due to instabilities in the accretiondisc causing the sudden collapse of large quantities of ma-terial onto the white dwarf. The nova-like variables (NLs)are the non-eruptive CVs, i.e. objects which have never been (cid:63)
E-mail: d.sahman@sheffield.ac.uk; vik.dhillon@sheffield.ac.uk observed to show nova or dwarf nova outbursts. The absenceof dwarf nova outbursts in NLs is believed to be due to theirhigh mass-transfer rates, producing ionised accretion discs inwhich the disc-instability mechanism that causes outburstsis suppressed (Osaki 1974). The mass transfer rates in NLsare ˙ M ∼ − − − M (cid:12) yr − whereas DNe have rates of˙ M ∼ − − − M (cid:12) yr − . Note that throughout this pa-per, when we refer to NLs we mean non-magnetic systems,i.e. we do not include in our definition systems that accretevia magnetic field lines, such as polars and intermediate po-lars.Our understanding of CV evolution has made greatstrides in recent years (e.g. see Knigge et al. 2011, Knigge2011). For example, there is now strong evidence that the2–3 hr period gap in the orbital period distribution of CVsis indeed due to disrupted angular momentum loss (Patter-son et al. 2005), the first “period-bounce” CVs, in whichthe secondary star has lost so much mass to the WD thatit falls below the hydrogen–burning limit, have now beendiscovered (Littlefair et al. 2008), and the predicted spike of c (cid:13) a r X i v : . [ a s t r o - ph . H E ] M a y D. I. Sahman et al systems at the period minimum has been revealed (G¨ansickeet al. 2009).One of the remaining unsolved problems in CV evolu-tion is: how can the different types of CV co-exist at the sameorbital period? Theory predicts that all CVs evolve fromlonger to shorter orbital periods on timescales of gigayears,and as they do so the mass-transfer rate also declines (e.g.see Fig. 19 of Knigge et al. 2011). At periods longer thanapproximately 5 hrs, all CVs should have high mass-transferrates and appear as NLs, whereas below this period the lowermass-transfer rate allows the disc-instability mechanism tooperate and all CVs should appear as DNe. This theoreticalexpectation, however, is in stark contrast to observations,which show that the fraction of nova-like variables to dwarfnovae is actually highest at 3 hr periods and then declinesto longer periods (e.g. see Fig. 18 of Knigge et al. 2011).It is possible that CVs cycle between NL and DNstates on timescales shorter than the gigayear evolutionarytimescale of the binary, thereby explaining the co-existenceof NLs and DNe at the same orbital period. Two mechanismsfor such a cycle have been proposed. Both mechanisms in-voke cyclical variation in the irradiation of the secondary,which in turn drives cyclical variation of ˙ M with timescalesof the order of ∼ − yrs.The first idea is that there is an irradiation feedbackmechanism. The flux from the WD illuminates the inner faceof the secondary which flattens the temperature gradient inthe photosphere, leading to an expansion in the radius ofthe secondary and an increase in ˙ M above the secular mean(see B¨uning & Ritter 2004 and references therein). The en-hanced ˙ M drives an increase in the radius of the secondary’sRoche lobe. Eventually the expansion of the secondary starcannot keep pace with the Roche lobe expansion, leading toa lower ˙ M , and hence a reduction in the irradiating flux.Consequently, the secondary begins to shrink and the feed-back mechanism operates in reverse as the mass transferrate reduces. B¨uning & Ritter (2004) found that this mech-anism could produce limit cycles in ˙ M of the appropriatetimescales (see Fig. 5 of Knigge et al. 2011), causing CVsto cycle between DN and NL states. However, their modelsshow that systems just above the period gap are actually sta-ble and do not undergo cycles. Hence, although this modelexplains why some NLs and DNe may co-exist at the sameorbital period, it does not explain why the irradiation-drivenfeedback mechanism would make the NL fraction highestaround 3 hr and decline towards longer periods.The second hypothesis for variable ˙ M is a nova-inducedcycle. Some fraction of the energy released in the nova eventwill heat up the WD, leading to increased irradiation andsubsequent bloating of the secondary. Following the novaevent, the system would have a high ˙ M and appear as a NL.As the WD cools, the radius of the secondary would returnto its secular value, and hence ˙ M will reduce and the systemchanges to a DN. In this model, therefore, CVs are expectedto cycle between nova, NL and DN states, on timescales of10 − yrs (see Shara et al. 1986), thereby explainingthe co-existence of these CV sub-types at the same orbitalperiod. Like the irradiation feedback mechanism, however,this nova cycle model does not explain why the NL fractionis highest around 3 hr and declines towards longer periods.The cyclical evolution of CVs through nova, NL andDN phases recently received observational support from the discovery that BK Lyn appears to have evolved through allthree phases since its likely nova outburst in the year AD 101(Patterson et al. 2013). A second piece of evidence has comefrom the discovery of nova shells around the dwarf novaeZ Cam and AT Cnc (Shara et al. 2007, Shara et al. 2012),verifying that they must have passed through an earlier novaphase. A more obvious place than DNe to find nova shellsis actually around NLs, as the nova-induced cycle theorysuggests that the high ˙ M in NLs could be due to a recentnova eruption. Finding shells around the highest accretion–rate NLs would lend further support to the existence of nova-induced cycles and hence help explain why systems withdifferent ˙ M are found at the same orbital period.In this paper we present the results of an H α imagingsurvey of the fields surrounding a sample of 101 cataclysmicvariables, 24 of them NLs, with the aim of identifying novashells around the central binary. The choices of telescope aperture and field of view for thisproject were dictated by the expected brightness and radiiof the nova shells around CVs. Recombination theory tellsus that the H α luminosity per unit volume of a nova shellis proportional to density squared, and hence the total lu-minosity of the shell is inversely proportional to volume. Ifwe assume that the shell expands at a constant velocity,then its volume increases as the cube of time. Therefore,the luminosity is inversely proportional to time cubed, andthe surface brightness decreases as time to the fifth power.This expectation has been empirically confirmed by Downeset al. (2001) who found that the H α surface brightness ofnova shells diminishes as t − . , although novae with strongshock interactions between the ejecta and any pre-existingcircumstellar material, e.g. GK Per (Shara et al. 2012) andT Pyx (Shara et al. 1989), do not fit this relationship.To estimate how bright the shells around NLs mightbe, we used as a guide the archetypal old nova DQ Her(Nova Her 1934) which has a clearly visible shell (Slavinet al. 1995). Its shell had a brightness of 1 . × − ergs/s/cm /arcsec and a radius of 8 (cid:48)(cid:48) in May 1982, 47.5years after its nova eruption (Ferland et al. 1984). This im-plies that 100 years after outburst it will fade to 2 . × − ergs/s/cm /arcsec , by which time its radius will have dou-bled to 16 (cid:48)(cid:48) , and at 200 years after outburst it will fade to7 . × − ergs/s/cm /arcsec , and its radius will havedoubled again to 32 (cid:48)(cid:48) . We can use these estimates to deter-mine our observing strategy and to establish the instrumen-tal setup that is required.Nova-like variables have apparent magnitudes of ∼ ∼
10 magnitudes during erup-tion, implying that most nova-like variables when in out-burst would have been just visible to the naked eye. Giventhe much increased rate of nova detection in the last 100years, it is very unlikely that such eruptions would havebeen missed if they had occurred within the last ∼
100 yearsor so. We thus expect to have to detect shells around nova-like variables which are at least 100 years old; any which c (cid:13) , 000–000 earching for nova shells erupted more recently than this would likely have been de-tected and would now be classified as a nova. Assuming theshell of DQ Her is representative, a 100-year old shell wouldbe at least 16 (cid:48)(cid:48) in radius and no brighter than 2 . × − ergs/s/cm /arcsec . Hence to detect such shells we requiredeep images with a relatively modest field of view, which ledus to use the Auxiliary Port on the 4.2m William HerschelTelescope (WHT) on La Palma (see Sec. 2.1.2).To estimate the age of the faintest shell we might detectwith this setup, we simulated the images we would obtainfrom a spherical nova shell in an 1800 second H α -exposure.The simulation showed that a shell with luminosity com-parable to the shell of DQ Her would become too faint todetect in an image at ∼
180 years after outburst. For approx-imately circular, small shells that are centred on the binary,this detection threshold can be pushed fainter by computingthe mean radial profile of the central object and inspectingthe wings for evidence of a shell (see Gill & O’Brien 1998).The DQ Her shell would be apparent in the radial profileup to 220 years after outburst. Assuming DQ Her is repre-sentative, this means that we would be able to detect shellsaround nova-like variables from nova eruptions up to a max-imum of ∼
220 years ago using our proposed setup on theWHT.
The observations were taken on the nights of 1997 Octo-ber 24–26. We used the 1024 × at the Cassegrain focus ofthe WHT to image the fields around our target nova-likevariables. This setup gave a platescale of 0 . (cid:48)(cid:48) per pixeland hence a field size of 113 (cid:48)(cid:48) × (cid:48)(cid:48) . H α is one of thestrongest features in the spectra of nova shells, with typi-cal velocity widths of up to 2000 km s − (e.g. Warner 1995).In order to maximise the detection of light from the shelland minimise the contribution of sky, we therefore used anarrow-band (55˚A FWHM = 2500 km s − ) interference fil-ter centred on the rest wavelength of H α (ING filter number61 ). Note that this filter also includes a contribution from[N ii ] 6584˚A emission, which may dominate the spectra ofnova shells with strong shock interaction of the ejecta withany pre-existing circumstellar medium, e.g. T Pyx (Sharaet al. 1989).As we were planning to compare the radial profiles ofthe target stars with field stars, we had to ensure that wedid not saturate the target stars. Hence the CCD chip wasused unbinned and in quick readout mode, in order to de-crease dead-time, at the expense of a negligible decrease insignal-to-noise (thanks to the fact that our observations werealways sky-limited). The observing conditions were excellentthroughout the run; the sky was always photometric, therewas no evidence of dust and the seeing was usually sub-arcsecond, with an occasional excursion up to 1.5–2 (cid:48)(cid:48) . http://catserver.ing.iac.es/filter/ To ensure we only targeted relatively well-studied systemswith reliable CV classifications, we made our selection fromthe catalogue of Ritter & Kolb (2003) (hereafter RK cata-logue). We selected a total of 31 CVs, predominantly NLs,and searched for nova shells around them. To test our setupwe included three systems with known nova shells, BT Mon,DQ Her and GK Per. We also took the opportunity to ob-serve two asynchronous polars, which are CVs with magneticWDs in which the spin period of the WD is not synchronisedwith the orbital period (Warner 1983). The asynchronicityis believed to be due to a recent nova event, as shown bythe system V1500 Cyg which had a nova eruption in 1975(Stockman et al. 1988). The two asynchronous polars weobserved with the WHT were V1432 Aql and BY Cam. Wealso included three other non-NL systems, PQ Gem, whichis an intermediate polar, IP Peg which is a DN, and AYPsc which is a Z Cam-type DN, which were favourably po-sitioned during our observing run.A full list of the 31 objects observed with the WHT anda journal of observations is given in Tab. 1. In summary, thetargets comprised 3 old novae with known shells, 1 old novawithout a known shell (V Per), 2 asynchronous polars, 1intermediate polar, 22 NLs, and 2 DNe. In Fig. 1 we showthe orbital period distribution of all the systems we observedwith the WHT compared to the total number of systems inthe RK catalogue. We deliberately selected a substantialnumber of the systems in the 3–4 hr orbital period range,which is where most NLs appear, as shown in Fig. 18 ofKnigge et al. (2011). The images were debiased using the median level of the over-scan strip and flat-fielded using normalised twilight sky flats.Where we had taken multiple images of targets, these werecombined to improve the signal-to-noise ratio. Sky subtrac-tion was performed by subtracting the median level deter-mined from two blank sky areas of size 100 ×
100 pixels.Each frame suffered from significant vignetting in the cor-ners due to the circular filter holder, which was not fullycorrected by the flat field. The corners of each image werehence removed by setting a series of 50 ×
50 pixel blocks toa fixed value, so that they appear white in the final imagesshown in Appendix A. Pixels affected by cosmic rays wereset to the average value of surrounding pixels. All processingwas performed using the kappa and figaro packages in the starlink suite of programs. In support of our WHT observations, we also examinedknown CVs in the 2.5m Isaac Newton Telescope (INT)Photometric H α Survey of the Northern Galactic Plane(IPHAS). IPHAS is a 1800 deg survey of the northern MilkyWay spanning the galactic latitude range − ◦ < b < +5 ◦ http://starlink.jach.hawaii.edu/starlinkc (cid:13)000
50 pixel blocks toa fixed value, so that they appear white in the final imagesshown in Appendix A. Pixels affected by cosmic rays wereset to the average value of surrounding pixels. All processingwas performed using the kappa and figaro packages in the starlink suite of programs. In support of our WHT observations, we also examinedknown CVs in the 2.5m Isaac Newton Telescope (INT)Photometric H α Survey of the Northern Galactic Plane(IPHAS). IPHAS is a 1800 deg survey of the northern MilkyWay spanning the galactic latitude range − ◦ < b < +5 ◦ http://starlink.jach.hawaii.edu/starlinkc (cid:13)000 , 000–000 D. I. Sahman et al
Table 1.
Journal of WHT observations. The classifications of the CVs have been taken from the RK catalogue. The date refers to thestart time of the first exposure. All shells detected in our WHT observations are shown in bold and discussed in Sect. 3.1. Note that theRK catalogue classifications for novae are N, Na, Nb. ∗ We detected a shell around V1315 Aql with the INT, not the WHT – see section3.2.5. † This system is a NL.Object Classification Orbital Date UTC UTC Number of Total exposure Visibleperiod (hrs) start end exposures time (secs) shell?PX And NL SW NS SH 3.51 25/10/97 00:06 01:11 2 3600 NUU Aqr NL UX SW SH 3.93 25/10/97 22:42 23:46 3 3300 NHL Aqr NL UX SW 3.25 27/10/97 00:13 00:54 2 2400 NV794 Aql NL VY 3.68 26/10/97 21:17 21:58 2 2400 NV1315 Aql NL UX SW 3.35 26/10/97 19:32 20:15 4 2400 N ∗ V1432 Aql NL AM AS 3.37 25/10/97 19:59 21:01 2 3600 NWX Ari NL UX SW 3.34 25/10/97 02:42 03:44 2 3600 NKR Aur NL VY NS 3.91 26/10/97 03:50 04:10 1 1200 NV363 Aur NL UX SW 7.71 25/10/97 05:09 06:10 2 3600 NBY Cam NL AM AS 3.36 25/10/97 04:00 05:01 2 3600 NAC Cnc NL UX SW 7.21 27/10/97 03:31 04:12 2 2400 NV425 Cas NL VY 3.59 24/10/97 22:29 22:37 2 200 NV751 Cyg † VY SW? NS SS 3.47 24/10/97 22:07 22:27 1 1200 NV1776 Cyg NL UX SW 3.95 25/10/97 21:31 22:32 2 3600 NCM Del NL UX VY? 3.89 26/10/97 20:22 21:05 4 2400 NPQ Gem NL IP 5.19 27/10/97 02:41 03:22 2 2400 N
DQ Her Na DQ 4.65 25/10/97 19:31 19:51 1 1200 Y
BH Lyn NL SW SH NS 3.74 26/10/97 04:24 05:05 2 2400 NBP Lyn NL UX SW 3.67 27/10/97 04:21 05:02 2 2400 N
BT Mon Na SW 8.01 26/10/97 06:12 06:43 2 1800 YBT Mon Na SW 8.01 27/10/97 06:02 06:32 1 1800 Y
V1193 Ori NL UX SW? 3.96 26/10/97 01:05 02:38 3 5400 NIP Peg DN UG 3.80 25/10/97 21:08 21:19 1 621 NLQ Peg NL VY SH NS 3.22 26/10/97 22:37 23:18 2 2400 NV Per Na NL SW? 2.57 27/10/97 01:06 01:48 2 2400 N
GK Per Na DN IP 47.92 25/10/97 06:14 06:34 1 1200 Y
AY Psc DN ZC NS 5.21 25/10/97 23:51 00:53 2 3600 NVY Scl NL VY 3.98 24/10/97 23:22 00:03 2 2400 NVZ Scl NL VY SW 3.47 24/10/97 22:47 23:39 2 3000 NSW Sex NL UX SW 3.24 26/10/97 05:15 05:56 2 2400 NRW Tri NL UX SW 5.57 25/10/97 01:29 02:30 2 3600 NDW UMa NL SW SH NS 3.28 27/10/97 05:09 05:50 2 2400 N and galactic longitude range 29 ◦ < l < ◦ . Three fil-ters were used, H α , Sloan r (cid:48) and Sloan i (cid:48) , reaching downto r (cid:48) ≈
20 (10 σ ). The survey took place between 2003 and2008. The survey used the INT Wide Field Camera (WFC)which offers a pixel scale of 0 . (cid:48)(cid:48) per pixel and a field ofview of ∼ (cid:48) × (cid:48) . Exposure times were initially set at120 s (H α ) and 10 s ( r (cid:48) and i (cid:48) ) but evaluation of the earlydata led to an increase in the r (cid:48) –band exposure time to 30 s– for full details of the observations and data reduction seeDrew et al. (2005) and Barentsen et al. (2014). We cross-matched the RK catalogue to the IPHAS footprint.There were 74 matches of CVs with the classification N, NLor DN (indicating nova, nova-like variable and dwarf nova,respectively). Each matching IPHAS field was reviewed vi-sually to determine whether any nebulosity was apparentaround the target CVs. Due to the significant H α nebulos- ity in the Galactic plane, we did not attempt to computeradial profiles for the IPHAS targets.The 74 systems we examined in IPHAS are listed inTab. 2. The targets comprised 2 asynchronous polars, 10polars & intermediate polars, 5 NLs, 34 DNe, 3 old novaewith known shells and 20 old novae without known shells.Three of the NLs, V1315 Aql, V363 Aur and V751 Cyg, werealso part of our WHT sample, as was BT Mon, an old novawith a known shell. In order to detect shells in the WHT images, we adoptedtwo strategies. First, we visually examined each image todetermine if a shell is visible. This technique would revealwide shells with diameters of more than a few arcseconds.Second, we calculated the radial profile of each CV and com-pared it to a number of field stars in the same image. Any c (cid:13) , 000–000 earching for nova shells Table 2.
List of CVs examined in the IPHAS database. The classifications of the CVs have been taken from the RK catalogue. Note thatthe RK catalogue classifications for novae are N, Na, Nb. All shells detected in IPHAS are shown in bold and are discussed in Sect. 3.2. ∗ We confirmed the detection of a shell around V1315 Aql with additional INT observations – see section 3.2.5. † This system is a NL. ‡ Whilst these objects are shown as possible Z Cam systems in the RK catalogue, Simonsen et al. (2014) found that they do not exhibitthe standstills necessary for this classification and that they are actually normal DNe.Object Classification Orbital Visible Object Classification Orbital Visibleperiod (hrs) shell? period (hrs) shell?CI Aql Nr 14.83 N V2468 Cyg Na 3.49 NKX Aql DN SU 1.45 N V2491 Cyg Na 2.56 NV368 Aql Na 16.57 N V446 Her Na DN 4.97 NV603 Aql Na SH NS 3.32 N CP Lac Na SW? 3.48 N
V1315 Aql NL UX SW 3.35 Y ∗ DI Lac Na 13.05 NV1425 Aql Na NL? IP? 6.14 N
BT Mon Na SW 8.01 Y
V1493 Aql Na 3.74 N CW Mon DN UP IP? 4.24 NV1494 Aql Na 3.23 N V902 Mon NL IP 8.16 NFS Aur DN UG IP PW? 1.43 N V959 Mon N 7.10 NHV Aur DN SU 1.98 N CZ Ori DN UG 5.25 N
T Aur Nb 4.91 Y
V344 Ori DN ZC ‡ ‡ V458 Vul Na 1.64 Y
EY Cyg DN UG SH? 1.10 N V498 Vul DN SU WZ 1.41 NV337 Cyg DN SU 1.64 N GD 552 DN? WZ? 1.71 NV503 Cyg DN SU NS 1.87 N Lanning 420 DN SU 1.45 NV516 Cyg DN UG 4.11 N J0130+6221 DN? 3.12 NV550 Cyg DN SU 1.62 N J0345+5335 CV DN? 7.53 NV751 Cyg † VY SW? NS SS 3.47 N J0506+3547 DN SU 1.62 NV1251 Cyg DN SU WZ 1.77 N J0518+2941 NL? 5.72 NV1316 Cyg DN SU 1.78 N J0524+4244 NL AM AS 2.62 N
V1363 Cyg DN ZC? ‡ ? J0619+1926 DN SU WZ 1.34 NV1454 Cyg DN SU 1.36 N J1853-0128 NL IP n/a N V1500 Cyg Na NL AM AS 3.35 Y
J1915+0719 DN SU WZ 1.37 NV2274 Cyg Na 7.20 N J1926+1322 NL IP 4.58 N
V2275 Cyg Na IP? 7.55 Y ? J1953+1859 DN SU? 1.44 NV2306 Cyg NL IP 4.35 N J2133+5107 NL IP 7.14 NV2362 Cyg Na 1.58 N J2138+5544 NL IP n/a NV2467 Cyg Na NL IP? 3.83 N J2250+5731 NL AM 2.90 N nebulosity around the CV due to a nova shell would causethe radial profile of the CV to lie above the average profile ofthe field stars (for example, see the radial profile of BT Monin Fig. A4). This technique can reveal shells with diametersof less than a few arcseconds, and was successfully used byGill & O’Brien (1998) to discover four new nova shells. A keyassumption in this technique is that the Point Spread Func-tion (PSF) is uniform across the WHT chip. Fig. 2 showsthe PSFs for five field stars (arrowed) in the image of V Per.The PSFs show identical radial profiles irrespective of fieldposition, giving confidence that the PSFs are uniform acrossthe field of view of the CCD, as expected.The centroids of the stars were first measured by fittinga two-dimensional Gaussian. The radial profiles were thengenerated by calculating the radial distance of each pixel from the centroid, and then averaging the fluxes of the pix-els falling into bins of increasing radial distance from thecentroid. The radial profiles were then normalised to unity,and plotted from the centre of the star until the flux reached1 σ above the mean background flux.In Appendix A we show the images and radial profilesfor all of the objects that were observed with the WHT. Asexpected, the images for the three old novae with previouslyknown shells (BT Mon, DQ Her, GK Per) clearly show ashell and each is discussed briefly below. There are no visibleshells in the images of the remaining objects, nor do any ofthe radial profiles differ significantly from the field stars. c (cid:13)000
V2275 Cyg Na IP? 7.55 Y ? J1953+1859 DN SU? 1.44 NV2306 Cyg NL IP 4.35 N J2133+5107 NL IP 7.14 NV2362 Cyg Na 1.58 N J2138+5544 NL IP n/a NV2467 Cyg Na NL IP? 3.83 N J2250+5731 NL AM 2.90 N nebulosity around the CV due to a nova shell would causethe radial profile of the CV to lie above the average profile ofthe field stars (for example, see the radial profile of BT Monin Fig. A4). This technique can reveal shells with diametersof less than a few arcseconds, and was successfully used byGill & O’Brien (1998) to discover four new nova shells. A keyassumption in this technique is that the Point Spread Func-tion (PSF) is uniform across the WHT chip. Fig. 2 showsthe PSFs for five field stars (arrowed) in the image of V Per.The PSFs show identical radial profiles irrespective of fieldposition, giving confidence that the PSFs are uniform acrossthe field of view of the CCD, as expected.The centroids of the stars were first measured by fittinga two-dimensional Gaussian. The radial profiles were thengenerated by calculating the radial distance of each pixel from the centroid, and then averaging the fluxes of the pix-els falling into bins of increasing radial distance from thecentroid. The radial profiles were then normalised to unity,and plotted from the centre of the star until the flux reached1 σ above the mean background flux.In Appendix A we show the images and radial profilesfor all of the objects that were observed with the WHT. Asexpected, the images for the three old novae with previouslyknown shells (BT Mon, DQ Her, GK Per) clearly show ashell and each is discussed briefly below. There are no visibleshells in the images of the remaining objects, nor do any ofthe radial profiles differ significantly from the field stars. c (cid:13)000 , 000–000 D. I. Sahman et al
Figure 1.
Orbital period distribution of the systems we observedwith the WHT (hatched) compared to the distribution of all theCVs in the RK catalogue. The left-hand dotted line indicates theperiod minimum and the central and right-hand dotted lines showthe period gap taken from Knigge et al. (2011).
Figure 2.
Left: WHT image of V Per (Star number 1). Right:PSFs for the five arrowed stars showing the uniformity across theWHT chip. Each numbered star has been plotted as follows 1 -solid line, 2 - dashed, 3 - dot-dashed-dot-dashed, 4 - dotted, 5 -dashed-dot-dot-dot-dashed. The orientation of the image is thesame as that shown in Fig. 3.
The shell around BT Mon (Nova Mon 1939) was discov-ered spectroscopically by Marsh et al. (1983). BT Mon isa high-inclination system and the system parameters werederived by Smith et al. (1998). The first image of the shellwas reported by Duerbeck (1987), who found it to be anincomplete clumpy, slightly elliptical ring with approximatedimensions of 11 (cid:48)(cid:48) × (cid:48)(cid:48) and the major axis pointing in theNW–SE direction.Our image and radial profile of BT Mon is shown inFig. A4. The lower right quadrant was not used to calculatethe radial profile in order to remove the flux from the nearbystar. The radial profile of BT Mon clearly deviates from theprofile of the field stars, from approximately 4 (cid:48)(cid:48) outwards.This is due to the presence of the shell and gives assurancethat our technique for identifying shells is valid.In Fig. 3 we show an enlarged version of our image ofthe BT Mon shell. We estimate the shell diameter to be13 (cid:48)(cid:48) ± (cid:48)(cid:48) . Assuming a constant shell expansion velocity of1800 ±
300 km s − and a distance of 1 . ± . Figure 3.
Enlarged image of the BT Mon nova shell. Notethat the second star to the lower right is an unassociated fore-ground/background star. by Marsh et al. (1983), together with the date of the novaas 1939 .
7, gives an expected diameter of 12 ± (cid:48)(cid:48) at the timeof our observations, in agreement with our measured value. DQ Her (Nova 1934) is an intermediate polar with systemparameters derived by Horne et al. (1993). The nova shell(see Fig. A4) is a prolate ellipsoid with a slightly pinchedcentral ring. Vaytet et al. (2007) used our WHT image of DQHer to estimate the angular size and hence distance of thesystem. They measured the angular size of the major andminor axes to be a = 25 . ± . (cid:48)(cid:48) and b = 18 . ± . (cid:48)(cid:48) .Assuming a constant expansion velocity of 370 ±
14 km s − ,they derived a distance of d = 525 ±
28 pc.
The nova GK Per (1901) is the archetypal nova remnantand has been extensively studied (see Shara et al. (2012)for a review). The shell is boxy in shape, of size approxi-mately 100 (cid:48)(cid:48) × (cid:48)(cid:48) and exhibits clumpy knots (see Fig. A5).Recently, Liimets et al. (2012) derived a three dimensionalmodel of the nova shell in GK Per, and determined theproper motion and radial velocities of more than 200 knots inthe ejecta. The knots have a wide range of velocities (600–1000 km s − ) and have suffered only modest deceleration.Shara et al. (2012) used HST images from 1995 and 1997 toresolve over 1000 filamentary structures in the ejecta. Theyalso investigated a jet-like feature, first discovered by Anu-pama & Prabhu (1993), which they suggest could be theshock interaction of a collimated flow with the ISM, proba-bly originating from the accretion disc. The jet extends some2.7 (cid:48) to the NW, which is larger than the field of view of ourimage. We examined our image of GK Per but could not c (cid:13) , 000–000 earching for nova shells find any evidence of the jet-like feature on smaller spatialscales, most probably due to the lower signal-to-noise ratioof our image. We visually examined the IPHAS images for evidence ofnova shells. Tab. 2 lists all of the objects we examined, andindicates whether a shell is visible. The short exposure timesof the H α images (120 s) means that only bright, nearbyshells are likely to be visible. We found three old novae withshells that are visible in the IPHAS footprint: T Aur (NovaAurigae 1891), V458 Vul (Nova Vul 2007 No. 1) and V1500Cyg (Nova Cygni 1975), all of which are well studied sys-tems. We briefly review these objects below. We found nodefinite detections of shells around any other IPHAS targetswith the exception of two systems, V1363 Cyg and V1315Aql, as discussed below. We did discover a nebula aroundV2275 Cyg, which is too large to be associated with its novaevent in 2001. This nebula may be a light echo due either toscattering off, or flash ionisation of, a pre-existing nebula.We also discuss this object further below. The IPHAS H α image of T Aur is shown in Fig. 4a. The shellis clearly discernible in the image giving us confidence thatit is possible to see nova shells in the IPHAS images. Theshell structure has been likened to that of DQ Her, althoughT Aur is some 43 years older (Slavin et al. 1995). The shellis elliptical in shape, with major and minor axes of length ∼ (cid:48)(cid:48) × (cid:48)(cid:48) respectively. The H α image of V458 Vul is shown in Fig. 4b. The shellhas major and minor axes of ≈ (cid:48)(cid:48) × (cid:48)(cid:48) . The image wastaken in June 2007, two months before the system underwenta nova explosion in August 2007. The shell is actually apre-existing planetary nebula ejected some 14,000 years ago(Wesson et al. 2008). The central binary is most likely apost-double common-envelope binary comprised of a WD ofmass ∼ (cid:12) and a post-AGB secondary of mass ∼ (cid:12) (Rodr´ıguez-Gil et al. 2010). V1500 Cyg (Nova Cygni 1975) is a well-studied nova andis the archetypal asynchronous polar (Wade et al. 1991).The nova shell was first imaged four years after outburstby Becker & Duerbeck (1980), who measured the radius at ∼ . (cid:48)(cid:48) . Subsequently, Wade et al. (1991) presented an im-age taken in 1987 by which time the shell had expanded to ∼ . (cid:48)(cid:48) , giving an expansion rate of 0 . (cid:48)(cid:48) per annum, andSlavin et al. (1995) presented an image taken in 1993 show-ing a nebular radius of ∼ (cid:48)(cid:48) . The IPHAS H α image takenin 2004 is shown in Fig. 4c. The nova shell is extremely faintand has a radius of ∼ (cid:48)(cid:48) , still consistent with the nebularexpansion rate of ∼ . (cid:48)(cid:48) per annum given by Wade et al.(1991). Figure 5. H α image of the nebula surrounding V1315 Aql. Theimage is 10 (cid:48) × (cid:48) , with North up and East to the left. V1315 Aqlis the bright star located at the centre of the image. In Fig. 4d we show the IPHAS H α image of the dwarf novaV1363 Cyg. The H α image is heavily populated with fieldstars making the surrounding nebula difficult to discern.Hence we show the H α − r (cid:48) image in Fig. 4e, which effec-tively removes most of the flux from the field stars. The fieldis extremely crowded and the object lies close to a ribbon ofgas, making the unambiguous detection of a nova shell ex-tremely difficult. However, there is a faint egg-shaped shellof emission of ≈ (cid:48) diameter, approximately centred on theCV. Figs. 4f & 4g show the IPHAS H α and H α − r (cid:48) images ofV1315 Aql. There is a faint shell of ≈ . (cid:48) radius approx-imately centred on the CV, with more pronounced emis-sion towards the West. We also imaged this object with theWHT. However, the small field of view of our WHT image(see Fig. A1) is not large enough to confirm the possibledetection of this shell.In order to confirm the detection of the shell aroundV1315 Aql, we took a further 13 exposures of V1315 Aqlon 2014 August 2 with the WFC on the INT with a totalexposure time of 7200 s in H α . The stacked H α image isshown in Fig. 5. The image clearly shows a shell surround-ing the central system, with a radius of ∼ (cid:48) , confirmingthe proposed detection in Fig. 4. Assuming a shell expan-sion rate of 2000 km s − and a distance of 356 +65 − pc (Aket al. 2007), this means that V1315 Aql experienced a novaeruption ∼
120 years ago.We examined the historic records of nova sightings com-piled by Ho (1962) and Stephenson (1976), but found noth-ing that coincides with the position of V1315 Aql. c (cid:13)000
120 years ago.We examined the historic records of nova sightings com-piled by Ho (1962) and Stephenson (1976), but found noth-ing that coincides with the position of V1315 Aql. c (cid:13)000 , 000–000 D. I. Sahman et al (a) T Aur, H α (b) V458 Vul, H α (c) V1500 Cyg, H α (d) V1363 Cyg, H α (e) V1363 Cyg, H α − r wide field (f) V1363 Cyg, H α − r (cid:48) zoom(g) V1315 Aql, H α (h) V1315 Aql, H α − r (cid:48) Figure 4.
IPHAS H α and H α − r (cid:48) images. In all images, North is up, East is left. The nova eruption of V2275 Cyg (Nova Cygni 2001 No. 2)occurred on 2001 August 19 (Nakamura et al. 2001). Thefield around V2275 Cyg was observed on five epochs duringthe IPHAS survey. The five images are shown in Figure 6.In the first three images, taken between November 2003and November 2006, a nebula of ≈ . (cid:48) diameter is clearlyapparent, but it has disappeared in the fourth and fifth im-ages taken in December 2008 and August 2009. Using theexpansion velocity and minimum distance derived by Kiss et al. (2002) of approximately 2000 km s − and 3 kpc respec-tively, the shell from the 2001 nova event should have beenno larger than 0 . (cid:48) by November 2006, which is the date ofthe last IPHAS image the nebula was visible in. Hence thenebula in the image can not be from the 2001 nova event.The most obvious explanation is that it is a light echo frommaterial ejected from the system by a previous event, suchas a nova shell or a planetary nebula. The angular diame-ter of the shell is 2 . (cid:48) ± . (cid:48) . Adopting the distance of 3–8kpc derived by Kiss et al. (2002) using maximum magnitude c (cid:13) , 000–000 earching for nova shells (a) November 2003 (b) November 2005 (c) November 2006(d) December 2008 (e) August 2009 Figure 6. H α images of V2275 Cyg from IPHAS. A faint nebula is apparent in images (a)–(c) but is not present in images (d) and (e).In all images, North is up, East is left. versus rate of decline relationships, the radius of the shell is3–12 × m. The time from the nova in 2001 to the dateof the first IPHAS image is 819 days giving a light radiusof 2 . × m. These two radii are broadly comparable, asexpected for a light echo. Assuming a typical nova shell ex-pansion velocity of 2000 km s − , the age of the shell can beestimated to be ∼
300 years. We have reviewed the litera-ture and can find no previous discussion of a nebula aroundV2275 Cyg. Indeed, the presence of light echoes around no-vae are relatively rare, and only GK Per, V732 Sgr, V458Vul and T Pyx have recorded echoes (Kapteyn 1901, Swope1940, Wesson et al. 2008, Sokoloski et al. 2013, respectively).There are three principal blobs of material that are ap-parent in the images, as highlighted in Figs 7 a–c. Blob Adoes not appear in Fig. 7(a) but appears in Figs. 7(b) &7(c). It appears to move southwards (towards the right inthe images) by approximately 20 (cid:48)(cid:48) . At a distance of 3 kpc,with an interval of approximately one year between the twoimages, this equates to a transverse speed of 2 . × ms − .Clearly this cannot be bulk motion of material. It is bet-ter explained as the passage of a light pulse through a pre-existing bi-polar nebula, with the axis of symmetry of thenebula pointing approximately perpendicular to the plane ofthe sky. This orientation is suggested by the lack of eclipsesin the light curve of V2275 Cyg (Balman et al. 2005). Blob Bappears in all three images and whilst different parts change intensity, there is no consistent motion shown. This blob ofmaterial is diagonally opposite blob A and could be the op-posite pole of a bi-polar nebula. Blob C only appears in Fig7(a).If the shell is due to a previous nova event this wouldmean that V2275 Cyg should be reclassified as a recurrentnova (RN), in agreement with Pagnotta & Schaefer (2014)who identified V2275 Cyg as a likely RN on the basis of itsoutburst light curve and spectrum. Our goal was to search for previously undetected nova shellsaround CVs, primarily nova-like variables. The results of oursearch are shown in Tab. 3.
We surveyed 22 NLs with the WHT and 5 NLs with IPHAS(three NLs were surveyed in both giving a total of 24 uniqueNLs), and found no shells with the WHT and evidence foronly one shell in IPHAS, V1315 Aql, which we subsequentlyconfirmed with additional INT observations (see Sect. 3.2.5).What can we deduce from our discovery of a shellaround one NL? Let us assume that all novae that occurred c (cid:13)000
We surveyed 22 NLs with the WHT and 5 NLs with IPHAS(three NLs were surveyed in both giving a total of 24 uniqueNLs), and found no shells with the WHT and evidence foronly one shell in IPHAS, V1315 Aql, which we subsequentlyconfirmed with additional INT observations (see Sect. 3.2.5).What can we deduce from our discovery of a shellaround one NL? Let us assume that all novae that occurred c (cid:13)000 , 000–000 D. I. Sahman et al (a) November 2003 (b) November 2005 (c) November 2006
Figure 7. H α images of V2275 Cyg from IPHAS. See text for a discussion about the three blobs of material. In all images, North is up,East is left. Table 3.
Summary of our search for nova shells.Nova-like Polars & Asynchronous Dwarf Old TotalVariables Intermediate Polars Novae NovaePolarsWHT Targets 22 1 2 2 4 31IPHAS Targets 5 10 2 34 23 74less: Duplicated objects –3 0 0 0 –1 –4Grand total of systems 24 11 4 36 26 101 in the last ∼
100 years would have been observed. Thesewould now be classified as old novae in the RK catalogue andhence would not appear in our sample of NLs. We also knowthat our observations are not sensitive to shells older than ∼
200 yrs (see Sect. 2.1.1). Hence our search for nova shellsaround NLs is only likely to find shells between 100 and 200years old. We found one shell in this 100-year window, outof 24 NLs surveyed, indicating that the lifetime of the NLphase lasts approximately 2400 yrs. This is consistent withthe order-of-magnitude estimate of 1,000 years derived byPatterson et al. (2013) for the NL phase for long-period CVs.Hence our results lend some support to the nova-induced cy-cle theory, although we are dealing with small number statis-tics; our survey of 24 NLs represents only 31% of the 78 NLsin the RK catalogue. We also acknowledge the incomplete-ness of our survey due to the small field of view of the WHTimages, and the shallow IPHAS images (see Sect. 4.3 for adiscussion). We also note that our IPHAS search included 7novae with previously discovered shells but we only found 4,and searches for shells around known novae tend to recovershells around only half of the targets (Downes & Duerbeck2000). With all of the assumptions, uncertainties and surveyinefficiencies detailed above, our estimate of the NL-phaselifetime should be viewed as a lower limit.As we were about to submit this paper, a paper ap-peared by Schmidtobreick et al. (2015), who presented theresults of a survey for nova shells around 10 DNe in the3–4 hr period range with low- ˙ M and 5 NLs that show lowstates (VY Scl stars). They found no shells, and used this to set a lower limit of 13 000 years on the nova recurrencetime. This is consistent with the lifetime of the NL phase of ∼ The WHT survey included two asynchronous polars, V1432Aql and BY Cam. The images and radial profiles of these twosystems are shown in Appendix A2. There are no traces ofnebulosity in either the images or the radial profiles of thesesystems. There are two other asynchronous polars in theIPHAS survey, J0524+4244 and V1500 Cyg, the archetypalasynchronous polar. There is no evidence of a shell aroundJ0524+4244. We did recover the previously known shellaround V1500 Cyg, originating from its nova eruption in1975 (see Sect. 3.2.3).Our results imply that we are unable to confirm thata recent nova eruption is the cause of the asynchronicity inthe white dwarf spin of these systems. However, it is perhapsnot surprising that we did not find any new shells given oursurvey limits and the synchronisation timescale. BY Cam,for example, is estimated to synchronise within ∼ c (cid:13) , 000–000 earching for nova shells In hindsight, our decision to use the old nova DQ Her as aguide for our WHT search strategy (Sect. 2.1.1) led us tounderestimate the optimal field of view for hunting for novashells. This is because DQ Her has relatively slow ejecta(350 km s − ; Warner 1995). The angular size of the shell isdetermined by the time since the nova eruption, the distanceto the CV, and the speed of the ejecta, and is given by thefollowing scaling relation: R ∼ (cid:48)(cid:48) t/
100 yr × v/ − d/ kpc , (1)where R is the angular radius of the shell in arcseconds, t is the time elapsed since the nova eruption, v is the shellexpansion velocity, and d is the distance to the CV. Hencea recent, distant nova with slow-moving ejecta ( t = 100 , v =500 , d = 2) would have a small shell of radius ∼ (cid:48)(cid:48) , whereasan older, nearby nova with fast moving ejecta ( t = 200 , v =2000 , d = 0 .
5) would have expanded to a radius of ∼ . (cid:48) .Hence the field of view of the Auxiliary port on the WHT( ∼ (cid:48) radius) was too small to detect such large shells. Thisis borne out by the size of the one shell that we did discoveraround V1315 Aql, which is ∼ . (cid:48) in radius, and the twoshells discovered by Shara et al. (2007; 2012) of radii 1 . (cid:48) (AT Cnc) and 15 (cid:48) (Z Cam). Another problem with havingsuch a small field of view is the paucity of field stars for theradial-profile technique (Sect. 3.1).The IPHAS survey, on the other hand, had more thanenough field of view ( ∼ (cid:48) radius per pointing) to dis-cover nova shells but suffered from very short exposure times(120 s), which we had no control over, and from being con-strained to the Galactic plane, making it difficult to pick outnova shells from the H α nebulosity. A more optimal surveyfor nova shells would have approximately the same field ofview as an IPHAS pointing, avoid the Galactic plane and beof similar depth to our WHT survey. We have performed an H α -imaging survey for nova shellsaround CVs. We imaged 31 CVs with the WHT, andsearched the IPHAS fields around 74 CVs.Our search focused on looking for shells around nova-like variables, as the nova-induced cycle theory suggests thatthese systems are most likely to have undergone a recentnova eruption. Of the 24 unique NLs we examined we foundevidence for only one shell around V1315 Aql, which has aradius of ∼ . (cid:48) , indicative of a nova eruption approximately120 years ago.The survey included 4 asynchronous polars, 2 observedwith the WHT (V1432 Aql and BY Cam) and 2 in IPHAS(J0524+4244 and V1500 Cyg) but we found no shells aroundany of them, except the previously known shell aroundV1500 Cyg. Hence we are unable to confirm whether theasynchronicity of the WD spin in these systems is due to arecent nova eruption.We find no unambiguous detections of nova shellsaround other classes of CV, but we did find tentative evi-dence of a faint shell around the dwarf nova V1363 Cyg (seeSect. 3.2.4). We also find evidence for a light echo aroundthe nova V2275 Cyg, which erupted in 2001, indicative of an earlier nova eruption ∼
300 years ago, thus making V2275Cyg a possible recurrent nova (see Sect. 3.2.6).
ACKNOWLEDGEMENTS
This paper makes use of data obtained as part of the INTPhotometric H α REFERENCES
Ak T., Bilir S., Ak S., Retter A., 2007, New Astronomy,12, 446Anupama G. C., Prabhu T. P., 1993, MNRAS, 263, 335Balman S¸., Yılmaz A., Retter A., Sayga¸c T., Esenoglu H.,2005, MNRAS, 356, 773Barentsen G., Farnhill H. J., Drew J. E., Gonz´alez-SolaresE. A., Greimel R., Irwin M. J., Miszalski B., Ruhland C.,Groot P., Mampaso A., Sale S. E., Henden A. A., Aung-werojwit A., Barlow M. J., Carter P. J., Corradi R. L. M.,et al., 2014, MNRAS, 444, 3230Becker H. J., Duerbeck H. W., 1980, PASP, 92, 792B¨uning A., Ritter H., 2004, A&A, 423, 281Campbell C. G., Schwope A. D., 1999, A&A, 343, 136Downes R. A., Duerbeck H. W., 2000, ApJ, 120, 2007Downes R. A., Duerbeck H. W., Delahodde C. E., 2001,Journal of Astronomical Data, 7, 6Drew J. E., Greimel R., Irwin M. J., Aungwerojwit A.,Barlow M. J., Corradi R. L. M., Drake J. J., G¨ansickeB. T., Groot P., Hales A., Hopewell E. C., Irwin J.,Knigge C., Leisy P., Lennon D. J., Mampaso A., MashederM. R. W. ., et al., 2005, MNRAS, 362, 753Duerbeck H. W., 1987, The ESO Messenger, 50, 8Ferland G. J., Williams R. E., Lambert D. L., Slovak M.,Gondhalekar P. M., Truran J. W., Shields G. A., 1984,ApJ, 281, 194G¨ansicke B. T., Dillon M., Southworth J., ThorstensenJ. R., Rodr´ıguez-Gil P., Aungwerojwit A., Marsh T. R.,Szkody P., Barros S. C. C., Casares J., de Martino D.,Groot P. J., Hakala P., Kolb U., Littlefair S. P., Mart´ınez-Pais I. G., et al., 2009, MNRAS, 397, 2170Gill C. D., O’Brien T. J., 1998, MNRAS, 300, 221Ho P. Y., 1962, Vistas in Astronomy, 5, 127 c (cid:13)000
Ak T., Bilir S., Ak S., Retter A., 2007, New Astronomy,12, 446Anupama G. C., Prabhu T. P., 1993, MNRAS, 263, 335Balman S¸., Yılmaz A., Retter A., Sayga¸c T., Esenoglu H.,2005, MNRAS, 356, 773Barentsen G., Farnhill H. J., Drew J. E., Gonz´alez-SolaresE. A., Greimel R., Irwin M. J., Miszalski B., Ruhland C.,Groot P., Mampaso A., Sale S. E., Henden A. A., Aung-werojwit A., Barlow M. J., Carter P. J., Corradi R. L. M.,et al., 2014, MNRAS, 444, 3230Becker H. J., Duerbeck H. W., 1980, PASP, 92, 792B¨uning A., Ritter H., 2004, A&A, 423, 281Campbell C. G., Schwope A. D., 1999, A&A, 343, 136Downes R. A., Duerbeck H. W., 2000, ApJ, 120, 2007Downes R. A., Duerbeck H. W., Delahodde C. E., 2001,Journal of Astronomical Data, 7, 6Drew J. E., Greimel R., Irwin M. J., Aungwerojwit A.,Barlow M. J., Corradi R. L. M., Drake J. J., G¨ansickeB. T., Groot P., Hales A., Hopewell E. C., Irwin J.,Knigge C., Leisy P., Lennon D. J., Mampaso A., MashederM. R. W. ., et al., 2005, MNRAS, 362, 753Duerbeck H. W., 1987, The ESO Messenger, 50, 8Ferland G. J., Williams R. E., Lambert D. L., Slovak M.,Gondhalekar P. M., Truran J. W., Shields G. A., 1984,ApJ, 281, 194G¨ansicke B. T., Dillon M., Southworth J., ThorstensenJ. R., Rodr´ıguez-Gil P., Aungwerojwit A., Marsh T. R.,Szkody P., Barros S. C. C., Casares J., de Martino D.,Groot P. J., Hakala P., Kolb U., Littlefair S. P., Mart´ınez-Pais I. G., et al., 2009, MNRAS, 397, 2170Gill C. D., O’Brien T. J., 1998, MNRAS, 300, 221Ho P. Y., 1962, Vistas in Astronomy, 5, 127 c (cid:13)000 , 000–000 D. I. Sahman et al
Horne K., Welsh W. F., Wade R. A., 1993, ApJ, 410, 357Kapteyn J. C., 1901, Astron. Nachr., 157, 201Kiss L. L., Gogh N., Vink´o J., Fur´esz G., Cs´ak B., DeBondH., Thomson J. R., Derekas A., 2002, A&A, 384, 982Knigge C., 2011, ArXiv e-prints, p. 1101.2901Knigge C., Baraffe I., Patterson J., 2011, ApJS, 194, 28Liimets T., Corradi R. L. M., Santander-Garc´ıa M.,Villaver E., Rodr´ıguez-Gil P., Verro K., Kolka I., 2012,ApJ, 761, 34Littlefair S. P., Dhillon V. S., Marsh T. R., G¨ansicke B. T.,Southworth J., Baraffe I., Watson C. A., Copperwheat C.,2008, MNRAS, 388, 1582Marsh T. R., Oke J. B., Wade R. A., 1983, MNRAS, 205,33PNakamura A., Tago A., Abe H., 2001, IAU Circ., 7686, 2Osaki Y., 1974, Pub. Astr. Soc. Japan, 26, 429Pagnotta A., Schaefer B. E., 2014, ApJ, 788, 164Patterson J., Kemp J., Harvey D. A., Fried R. E., Rea R.,Monard B., Cook L. M., Skillman D. R., Vanmunster T.,Bolt G., Armstrong E., McCormick J., Krajci T., JensenL., Gunn J., Butterworth N., Foote J., Bos M., Masi G.,Warhurst P., 2005, PASP, 117, 1204Patterson J., Uthas H., Kemp J., de Miguel E., Krajci T.,Foote J., Hambsch F.-J., Campbell T., Roberts G., CejudoD., Dvorak S., Vanmunster T., Koff R., Skillman D., Har-vey D., Martin B., Rock J., Boyd D., Oksanen A., MorelleE., Ulowetz J., Kroes A., Sabo R., Jensen L., 2013, MN-RAS, 434, 1902Ritter H., Kolb U., 2003, A&A, 404, 301, (Update RKCat7.20, 2013)Rodr´ıguez-Gil P., Santander-Garc´ıa M., Knigge C., CorradiR. L. M., G¨ansicke B. T., Barlow M. J., Drake J. J., DrewJ., Miszalski B., Napiwotzki R., Steeghs D., Wesson R.,Zijlstra A. A., Jones D., Liimets T., Others 2010, MNRAS,407, L21Schmidtobreick L., Shara M., Tappert C., Bayo A., Edero-clite A., 2015, ArXiv e-prints, p. 1502.05230Shara M. M., Livio M., Moffat A. F. J., Orio M., 1986,ApJ, 311, 163Shara M. M., Martin C. D., Seibert M., Rich R. M., SalimS., Reitzel D., Schiminovich D., Deliyannis C. P., Sar-razine A. R., Kulkarni S. R., Ofek E. O., Brosch N., L´epineS., Zurek D., De Marco O., Jacoby G., 2007, Nat, 446, 159Shara M. M., Mizusawa T., Wehinger P., Zurek D., MartinC. D., Neill J. D., Forster K., Seibert M., 2012, ApJ, 758,121Shara M. M., Potter M., Williams R. E., Cohen J., 1989,ApJ, 337, 720Shara M. M., Zurek D., De Marco O., Mizusawa T.,Williams R., Livio M., 2012, AJ, 143, 143Simonsen M., Boyd D., Goff W., Krajci T., Menzies K.,Otero S., Padovan S., Poyner G., Roe J., Sabo R., SjobergG., Staels B., Stubbings R., Toone J., Wils P., 2014, Jour-nal of the American Association of Variable Star Ob-servers (JAAVSO), 42, 177Slavin A. J., O’Brien T. J., Dunlop J. S., 1995, MNRAS,276, 353Smith D. A., Dhillon V. S., Marsh T. R., 1998, MNRAS,296, 465Sokoloski J. L., Crotts A. P. S., Lawrence S., Uthas H.,2013, ApJ, 770, L33Stephenson F. R., 1976, QJRAS, 17, 121 Stockman H. S., Schmidt G. D., Lamb D. Q., 1988, ApJ,332, 282Swope H. H., 1940, Harvard College Observatory Bulletin,913, 11Vaytet N. M. H., O’Brien T. J., Rushton A. P., 2007, MN-RAS, 380, 175Wade R. A., Ciardullo R., Jacoby G. H., Sharp N. A., 1991,AJ, 102, 1738Warner B., 1983, in Livio M., Shaviv G., eds, CataclysmicVariables and Related Physics IAU Colloquia No. 72.p. 155Warner B., 1995, Cataclysmic Variable Stars. CambridgeUniversity Press, CambridgeWesson R., Barlow M. J., Corradi R. L. M., Drew J. E.,Groot P. J., Knigge C., Steeghs D., Gaensicke B. T., Napi-wotzki R., Rodriguez-Gil P., Zijlstra A. A., Bode M. F.,Drake J. J., Frew D. J., Gonzalez-Solares E. A., et al.,2008, ApJ, 688, L21
APPENDIX A: WHT IMAGES (INALPHABETICAL ORDER OFCONSTELLATION) c (cid:13) , 000–000 earching for nova shells Figure A1.
WHT images of our target CVs (left) and the as-sociated radial profiles (right); the solid line is the radial profileof the CV and the dashed lines are field stars. The radial profilesare normalised to unity and plotted until they reach 1 σ above thebackground level. The CVs are marked by bars and are locatedtowards the centres of the images. The orientation of all imagesis the same, and is shown in the image of PX And. Figure A2.
See caption to Figure A1 for details.c (cid:13)000
See caption to Figure A1 for details.c (cid:13)000 , 000–000 D. I. Sahman et al
Figure A3.
See caption to Figure A1 for details.
Figure A4.
See caption to Figure A1 for details.c (cid:13) , 000–000 earching for nova shells Figure A5.
See caption to Figure A1 for details.
Figure A6.
See caption to Figure A1 for details.c (cid:13)000