Non-Detection of Nova Shells Around Asynchronous Polars
MMNRAS , 1–4 (2016) Preprint 1 October 2018 Compiled using MNRAS L A TEX style file v3.0
Non-Detection of Nova Shells Around Asynchronous Polars
Ashley Pagnotta (cid:63) and David Zurek Department of Astrophysics, American Museum of Natural History, New York, NY 10024
Accepted 2016 February 19. Received 2016 February 5; in original form 2016 February 5
ABSTRACT
Asynchronous polars (APs) are accreting white dwarfs (WDs) that have di ff erent WD and or-bital angular velocities, unlike the rest of the known polars, which rotate synchronously (i.e.,their WD and orbital angular velocities are the same). Past nova eruptions are the predictedcause of the asynchronicity, in part due to the fact that one of the APs, V1500 Cyg, was ob-served to undergo a nova eruption in 1975. We used the Southern African Large Telescope10m class telescope and the MDM 2.4m Hiltner telescope to search for nova shells aroundthree of the remaining four APs (V1432 Aql, BY Cam, and CD Ind) as well as one Inter-mediate Polar with a high asynchronicity (EX Hya). We found no evidence of nova shells inany of our images. We therefore cannot say that any of the systems besides V1500 Cyg hadnova eruptions, but because not all post-nova systems have detectable shells, we also cannotexclude the possibility of a nova eruption occurring in any of these systems and knocking therotation out of sync. Key words: novae, cataclysmic variables
Polars are accreting white dwarf (WD) binaries characterized bythe presence of a strong magnetic field ( ∼ −
230 MG; Ferrarioet al. 2015), which prevents the formation of an accretion disc andinstead channels the accreted material directly on to the poles of theWD. In most polars, the measured WD and orbital angular veloci-ties are found to be identical. For a handful of systems, however, theWD rotates asynchronously, so that those two velocity values di ff ersignificantly. The five known asynchronous polars (APs) are V1432Aql (RXJ 1940-10), BY Cam, V1500 Cyg, CD Ind (RXJ 2115-58),and Paloma (RX J0524 +
42) (Campbell & Schwope 1999; Schwarzet al. 2004). EX Hya is an Intermediate Polar (IP), with a magneticfield of ∼ / pressure and igniting a thermonu-clear runaway in the accreted layer, the ejection of which causesthe WD to spin with a higher angular velocity than before the nova(Campbell & Schwope 1999). V1500 Cyg (Nova Cyg 1975) wasobserved to be an AP after its nova eruption, lending support tothis theory, although its pre-eruption status is unknown. BY Cam,V1500 Cyg, and CD Ind all have a positive ω/ Ω , where ω is thesynodic angular velocity of the WD primary and Ω is the orbitalangular velocity. ω/ Ω is then a measurement of the asynchronicityof the system. Another way to measure asynchronicity is by looking (cid:63) E-mail: [email protected] at the percent di ff erence between the periods, defined as P orb − P spin P orb ,where P orb is the orbital period of the binary system and P spin is thespin period of the WD. The percent di ff erence and ω/ Ω , along withother general properties, are listed in Table 1 for each system. BYCam, V1500 Cyg, CD Ind, and Paloma all have positive percentdi ff erences, although they range over approximately one and a halforders of magnitude. V1432 Aql, however, is under-synchronous,with a negative ω/ Ω and percent di ff erence, which may indicate adi ff erent formation mechanism, although at this point the details ofthe theory are poorly understood in general, and particularly whenit comes to explaining how to obtain an under-synchronous AP.These systems do not remain asynchronous indefinitely; in-stead, they likely start returning to a synchronous state quickly af-ter being knocked out of sync. Models of this process vary, leadingto a range of estimates for the time needed to return to synchro-nization ( t sync ) for each system, also listed in Table 1. BY Camhas the largest range, with t sync estimates ranging from 250 ± > t sync estimated to be just a few hundreds of years formost APs, and the postulate that the asynchronicity was originallycaused by a nova eruption, it is reasonable to search for nova shellsaround the systems and expect to find something. Detection of sucha shell would eventually allow for an estimate of the date of thenova that caused the asynchronicity and thus provide another con-straint on the resynchronization time-scales as well as a further clueto the cause of the asynchronicity in the first place. Sahman et al.(2015) searched for shells around just two of our targets, V1432Aql and BY Cam, using the Auxiliary Port on the 4.2 m WilliamHerschel Telescope on La Palma and did not find evidence for anyshells. c (cid:13) a r X i v : . [ a s t r o - ph . S R ] M a r Pagnotta & Zurek
We observed two APs, V1432 Aql and CD Ind, and the IP EX Hya,in H α ( λ peak = . =
10 nm) using the SALTI-CAM imager on the 10m class South African Large Telescope(SALT) located at the South African Astronomical Observatory,near Sutherland, South Africa (Buckley et al. 2006; O’Donoghueet al. 2006). V1432 Aql was observed for a total of 3120s, dividedevenly between 26 di ff erent exposures and two nights (2013 June29 and 2013 July 11). With the same filter, CD Ind was observedon 2013 June 28 for a total of 1560s across 13 exposures, and EXHya on 2013 July 11 for 1800s spread over 15 di ff erent exposures.(There is e ff ectively no guide camera that can be used with SALTI-CAM, hence the large number of short exposures.) Additionally,we observed BY Cam in H α ( λ peak = . = (cid:48) unvignetted diameter for OSMOS as opposed tothe 8 (cid:48) diameter field of view of SALTICAM.For all targets, the usable individual images were processedusing the usual reduction steps in PyRAF and then summed using imcombine to obtain the greatest possible signal for each target.Compared to the observations from Sahman et al. (2015), we areable to go deeper on V1432 Aql and search a larger field of viewfor both V1432 Aql and BY Cam. The left-side images in Figures1 to 4 show the stacked images for each system observed; no in-dications of any shells can be seen. To highlight any faint edges inthe images and thus more thoroughly search for shells or shell frag-ments, we unsharp masked each combined image. The unsharpenedversions are shown in the right side images of Figures 1 to 4. Noneof the systems observed show visual evidence for a nova shell atany observed distance from the target star. Following a nova eruption, the ejected shell will expand until it dis-sipates into the circumstellar medium. From Table 3 of Pagnotta &Schaefer (2014), which collects observed nova characteristics, theaverage expansion velocity of a classical nova is 1800 km s − . As-suming this expansion velocity for any prior nova eruptions in thesystems we observed, and considering the resolution of the detec-tors and the site conditions, we can calculate the size and age of allpossible detectable shells. The distances used for each calculationare listed in Table 1. A certain number of years after an eruption,the shell will have expanded enough that it will be distinct fromthe image of the nova itself on the image (accounting for seeing,binning, and instrument e ff ects), and first detectable in its smalleststate. For each system, we assume the shell must be at least 5 pixelsaway from the star to be seen on the image; the minimum detectableshell sizes and ages are listed in Table 2. For all of the systems weobserved, shells from very recent novae can be expected to be seen,within the past two years for most of them. As the shell expands,eventually it will reach the edge of the field of view of the CCD.Since no dithering patterns were employed in our observations, weassume the total observed field corresponds to the full, unvignettedsize of the instrument / chip, with the target system located at thecentre. Using triangle geometry, we can calculate the physical shellsize and thus the possible age, which then gives us a date back to which we can assume we may have seen a shell if the nova haderupted that recently and produced an observable shell.For each system we calculate the shell sizes and ages assuming(a) constant expansion velocity, and (b) an expansion velocity thatdecelerates as described in Duerbeck (1987), which gives a meanhalf-lifetime of the velocities of 75 yr. From the Duerbeck paper,we can construct an exponential decay equation for the velocityover time: v ( t ) = . × (cid:32) (cid:33) t km yr − . (1)Taking the indefinite integral of Equation 1 and solving for the in-tegration constant given that r ( t = = r ( t ) = − . × e − . · t + . × km . (2)To calculate how long it would take the decelerating shell to reachthe edge of our fields of view, we need time as a function of radius,so we rearrange Equation 2 to obtain t ( r ) = − .
009 ln (cid:18) − r . × (cid:19) yr . (3)The amount of time necessary to observe the shells at their small-est sizes is the same for both cases (a) and (b), because it is such ashort amount of time that no significant deceleration can have oc-curred on a level that we would be able to detect. Deceleration,however, does change the largest possible shell ages, increasingthe amount of time we can expect to see the shell after the erup-tion, or essentially how far back in time we would be able to de-tect an eruption, because it takes longer for the shell to expand be-yond the field of view if it is decelerating. If we take the Duerbeck(1987) formulation at face value, we notice that the radius of theshell has an asymptote at r = . × km. In some cases, thefields of view of our images are larger than this, so theoretically wecould say that we would see all possible nova shells from an infiniteamount of time in the past, however this is clearly unphysical, be-cause we have observed at least two nova shells that are larger thanthe r = . × km limit. AT Cnc and Z Cam, two dwarf novaewith ancient nova shells (Shara et al. 2007, 2012a,b), have shellswith measured radii of 6 . × km and 2 . × km. TheDuerbeck (1987) result was empirically determined using observa-tions of just four novae, so it is not altogether surprising that it isnot universally applicable, but nevertheless it is a good first-orderapproximation of what we can expect, at least for the first 75 yrafter a nova eruption. There is likely a strong dependence on the lo-cal circum- and inter-stellar medium, but measuring and modellingthat is beyond the scope of this paper. For the cases in which ourfields of view are larger than the r = . × km asymptote,we can say only that we can see shells further back in time than inthe no deceleration case, but cannot put a firm upper limit on thetimeframe.For V1432 Aql, we can rule out nova shells from eruptionsthat happened up to 118 or 145 yr ago in the constant expansionvelocity case, depending on which distance measurement we use(187 or 230 pc, respectively; Ak et al. 2008; Barlow et al. 2006).Accounting for the deceleration of the shell, for both distances, wehave cases where the field of view is larger than the asymptote,so we can say that 118 and 145 yr are lower limits. For CD Ind,there are no shells detected from eruptions up to 59 yr ago in theconstant velocity case, and 86 yr ago with a decelerating shell. EXHya, the closest of our systems and the only non-AP, does not showshells from eruptions up to 35 yr ago or 43 yr ago, for constant anddecelerating shell velocities, respectively. For BY Cam, with the MNRAS , 1–4 (2016) on-Detection of Nova Shells Around Asynchronous Polars caveat that the image is shallow, in the constant expansion velocitysituation we rule out shells from eruptions that occurred as far backas 82 to 300 yr ago, again depending on the distance adopted (52 or190 pc; Ak et al. 2008; Barlow et al. 2006). If the shell deceleratesas described above, we can rule out shells from eruptions datingback to anywhere from 154 to <
300 yr ago. These time constraintsare listed for each system in Table 2.Additionally, we were able to check whether the APs havelarge-scale ultraviolet-bright shells, similar to those found aroundZ Cam (Shara et al. 2007) and AT Cancri (Shara et al. 2012a). Wesearched the GALEX archive and found all but V1500 Cyg haveimages in both the FUV and NUV bands (135.0-175.0 nm and175.0-280.0 nm, respectively). No shell is visible on any scale forany of the targets. The point spread function of GALEX is ∼ (cid:48)(cid:48) andconfirms the H α non-detections on scales larger than this.It is critical to remember that the lack of a nova shell does notequate to the lack of a nova eruption. Wade (1990) provides oneof the first statistical looks at how many shells have been detectedaround classical novae, reporting that 26 of the approximately 200known at the time had resolved remnants. Downes & Duerbeck(2000) did a survey of 30 recent, relatively nearby novae and found14 shells using a combination of ground- and space-based imaging,giving a 47% detection rate, although we note that this may be dif-ferent from the actual shell formation rate. There are many reasonsa shell may not be observed around a nova after its eruption even ifit has formed: the amount of mass ejected might be so small that theshell density is low and the shell is undetectable even shortly afterthe eruption; or, enough time has passed since the eruption that, asthe shell has expanded, its density and therefore surface brightnesshave decreased, making it too faint to detect; or, the shell mighthave expanded so quickly that it is larger than the field of view ofthe image.There is another possibility for finding old nova eruptions inthese systems: one can check through the major astronomical platearchives, namely those at the Harvard College Observatory in Cam-bridge, MA, and the Sonneberg Observatory in Sonneberg, Ger-many. There are scanning operations underway at both archives,which will allow for a quick check of the past behaviour of each ofthese objects once the fields are scanned and released. Although theHarvard operation, DASCH (Grindlay et al. 2012), has entered pro-duction scanning mode, it will likely be at least a few more yearsbefore all of the AP fields are scanned and available. Sonnebergis also scanning its plates, however they are not readily accessi-ble o ff site. Additionally, for a fully complete eruption search, it isrecommended that the plates be examined by hand for evidence oferuption, especially in crowded fields, because it is possible that theeruption is only captured on one plate, and if for whatever reasonthat plate is not properly solved by the software pipeline, it willnot be included in the digitized light curve results and the eruptionwill be missed. Although this method of searching for eruptions inarchival plates only covers the last ∼
120 yr, it is still a valuable re-source in the attempt to find previous eruptions, and allows for thepossibility of finding nova eruptions in systems that did not formdetectable shells.With no shell detections in our images, we cannot prove thatany of the systems in our study—the three APs V1432 Aql, BYCam, and CD Ind, and the IP EX Hya—had nova eruptions in therecent past that caused their asynchronicity today; we also cannotconclude that they did not have nova eruptions. It is possible thatthey erupted recently and the shells are fainter than our observa-tions, for any of the possible reasons discussed above. In this case,deeper imaging, especially for BY Cam, is advised. It is also pos- sible that the eruptions were further in the past than expected (i.e.our understanding of the models used to obtain t sync are incorrect),especially for BY Cam, with its possible synchronization time of > ACKNOWLEDGEMENTS
This research was supported by the Kathryn W. Davis PostdoctoralScholar programme, which is supported in part by the New YorkState Education Department and by the National Science Founda-tion under grant numbers DRL-1119444 and DUE-1340006.This manuscript is based on observations made with theSouthern African Large Telescope (SALT) under program 2013-1-AMNH-004 (PI: A. Pagnotta). We gratefully acknowledge thatAMNH access to SALT is made possible by a generous donationfrom the late Paul Newman and the Newman Foundation.This work is also based on observations obtained at the MDMObservatory, operated by Dartmouth College, Columbia Univer-sity, Ohio State University, Ohio University, and the University ofMichigan, using a filter borrowed from KPNO / NOAO, which washelpfully arranged by Eric Galayada at MDM. PyRAF is a prod-uct of the Space Telescope Science Institute, which is operated byAURA for NASA.We thank Tom Maccarone for the initial discussion thatsparked the idea for this project, Mike Shara for helpful discus-sions on the subject of nova shells, and Arlin Crotts for access tohis MDM time. Jana Grcevich provided many useful suggestionsthroughout the course of this work, and Denise Revello providedinvaluable coaching via the RAISE-W program; we are grateful tothem both. The writing of this manuscript was continually accom-panied by the dulcet tones of NPG, FH, the rest of the TBS Crew,and The Nixtape, which undoubtedly contributed to increased pro-ductivity.
REFERENCES
Ak, T., Bilir, S., Ak, S., & Eker, Z. 2008, New Astron., 13, 133Andronov, I. L., & Baklanov, A. V. 2007, Astrophysics, 50, 105Barlow, E. J., Knigge, C., Bird, A. J., J Dean, A., Clark, D. J., Hill, A. B.,Molina, M., & Sguera, V. 2006, MNRAS, 372, 224Buckley, D. A. H., Swart, G. P., & Meiring, J. G. 2006, in Society ofPhoto-Optical Instrumentation Engineers (SPIE) Conference Series,Vol. 6267, Society of Photo-Optical Instrumentation Engineers (SPIE)Conference Series, 0Campbell, C. G., & Schwope, A. D. 1999, A&A, 343, 132Downes, R. A., & Duerbeck, H. W. 2000, AJ, 120, 2007Duerbeck, H. W. 1987, Ap&SS, 131, 461Ferrario, L., de Martino, D., & Gänsicke, B. T. 2015, Space Sci. Rev.Grindlay, J., Tang, S., Los, E., & Servillat, M. 2012, in IAU Symposium,Vol. 285, IAU Symposium, ed. E. Gri ffi n, R. Hanisch, & R. Seaman,29–34MNRAS000
Ak, T., Bilir, S., Ak, S., & Eker, Z. 2008, New Astron., 13, 133Andronov, I. L., & Baklanov, A. V. 2007, Astrophysics, 50, 105Barlow, E. J., Knigge, C., Bird, A. J., J Dean, A., Clark, D. J., Hill, A. B.,Molina, M., & Sguera, V. 2006, MNRAS, 372, 224Buckley, D. A. H., Swart, G. P., & Meiring, J. G. 2006, in Society ofPhoto-Optical Instrumentation Engineers (SPIE) Conference Series,Vol. 6267, Society of Photo-Optical Instrumentation Engineers (SPIE)Conference Series, 0Campbell, C. G., & Schwope, A. D. 1999, A&A, 343, 132Downes, R. A., & Duerbeck, H. W. 2000, AJ, 120, 2007Duerbeck, H. W. 1987, Ap&SS, 131, 461Ferrario, L., de Martino, D., & Gänsicke, B. T. 2015, Space Sci. Rev.Grindlay, J., Tang, S., Los, E., & Servillat, M. 2012, in IAU Symposium,Vol. 285, IAU Symposium, ed. E. Gri ffi n, R. Hanisch, & R. Seaman,29–34MNRAS000 , 1–4 (2016) Pagnotta & Zurek
Table 1.
The Asynchronous PolarsSystem P orb (h) P spin (h) P orb − P spin P orb ω/ Ω t sync (yr) Distance (pc)V1432 Aql 3.3655 [1] − . × − [1] 110; 199 + − ; 96 . ± . . × − [1] 1200; 1107; ≥ ±
20 [7; 1; 8; 9] 52; 190 [5; 6]V1500 Cyg 3.3507 [1] 3.2917 [2] 0.0176 1 . × − [1] 185; 150; 150-290 [7; 1; 9] 1038 [5]EX Hya . × − [1] . . . 94 [5]Paloma 2.6195 [12] 2.4328; 2.2709 [12] 0.0713; 0.1331 . . . . . . ≤
240 [12]
Table 2.
Shell Size & Age LimitsSystem Smallest Shell Size ( (cid:48)(cid:48) ) Smallest Shell Age (yr) Largest Shell Age (yr) Largest Shell Age (yr) V1432 Aql > > > References: [1] Campbell & Schwope (1999); [2] Ramsay et al. (1999);[3] Staubert et al. (2003); [4] Andronov & Baklanov (2007); [5] Ak et al.(2008); [6] Barlow et al. (2006); [7] Piirola et al. (1994); [8] Honeycutt &Kafka (2005); [9] Pavlenko et al. (2013); [10] Hellier & Sproats (1992);[11] Hellier (1996); [12] Schwarz et al. (2007) EX Hya is an IP. This paper has been typeset from a TEX / L A TEX file prepared by the author. Assuming a constant expansion velocity of the nova shell after the erup-tion. Assuming a nova shell that decelerates after the eruption due to inter-actions with the circum- and inter-stellar medium. See Section 3 for moreinformation on how the deceleration was calculated, as well as details onthe " > " values in this column. V1432 Aql has two values in the Smallest and Largest Shell Age columnsdue to the two distances reported in the literature. The smaller values corre-spond to the 187 pc distance from Ak et al. (2008) and the larger to the 230pc distance from Barlow et al. (2006). BY Cam also has two values in the Smallest and Largest Shell Agecolumns due to two reported distances of 52 pc (Ak et al. 2008) and 190pc (Barlow et al. 2006). MNRAS , 1–4 (2016) on-Detection of Nova Shells Around Asynchronous Polars Figure 1.
Left:
SALTICAM H α image of the AP V1432 Aql (RXJ 1940-10), made from a combination of 26 separate exposures taken over the course of twonights, 2013 June 29 and 2013 July 11, for a total exposure time of 3120s. North is up, East is to the left, and the length of the directional arrows correspondsto 1 (cid:48) on the figure. The full unvignetted field of view of the image is 8 (cid:48) in diameter, and the lighter stripes in the middle of the image are due to the gap betweenthe two SALTICAM chips. The position of V1432 Aql is marked, and no shells or shell fragments are visible. Right:
The right side of this figure shows thesame V1432 Aql field seen on the left after it has been processed using an unsharp masking technique that involves subtracting a duplicate of the image fromthe original, after the duplicate has been shifted by 0.5 pixels in the x-direction. This method increases the local contrast of the di ff erent areas of the image andhighlights edge features, such as those seen in nova shells. With this, we can be confident that if there were a detectable shell in this image, we would see ithere.MNRAS000
The right side of this figure shows thesame V1432 Aql field seen on the left after it has been processed using an unsharp masking technique that involves subtracting a duplicate of the image fromthe original, after the duplicate has been shifted by 0.5 pixels in the x-direction. This method increases the local contrast of the di ff erent areas of the image andhighlights edge features, such as those seen in nova shells. With this, we can be confident that if there were a detectable shell in this image, we would see ithere.MNRAS000 , 1–4 (2016) Pagnotta & Zurek
Figure 2.
Left:
MDM 2.4m OSMOS H α image of the AP BY Cam, constructed from two images taken on 2015 March 15, for a total of 2400s of exposuretime. North is up, East is to the left, and the length of the directional arrows corresponds to 1 (cid:48) on the figure. (The larger field of view of OSMOS compared toSALTICAM is immediately visible by comparing the relative sizes of the 1 (cid:48) lines between this figure and those in Figures 1, 3, and 4.) The full field of viewof the image is 20 (cid:48) in diameter, and the large light spot in the eastern half of the image is due to the guide camera blocking part of the frame. The position ofBY Cam is marked, and no shells or shell fragments are visible. Right:
This figure shows the unsharp mask technique applied to the image of BY Cam on theleft side, using the same procedure described in the caption to Figure 1. Again, no shells or shell fragments are visible in the image.
Figure 3.
Left:
SALTICAM H α image of the IP EX Hya, made from a combination of 15 separate exposures taken on 2013 July 11, for a total exposure timeof 1800s. North is up, East is to the left, and the length of the directional arrows corresponds to 1 (cid:48) on the figure. The full field of view is 8 (cid:48) in diameter, andthe lighter stripe in the middle of the image is due to the gap between the two SALTICAM chips. The position of EX Hya is marked on the image, and thereare no shells or shell fragments visible. Right:
The unsharp masked image of the intermediate polar EX Hya, again processed using the same steps describedin the caption for Figure 1, also does not show any shells or shell fragments. MNRAS , 1–4 (2016) on-Detection of Nova Shells Around Asynchronous Polars Figure 4.
Left:
SALTICAM H α image of the AP CD Ind (RXJ 2115-58), made from a combination of 13 separate exposures taken on 2013 June 28, for atotal exposure time of 1560s. North is up, East is to the left, and the length of the directional arrows corresponds to 1 (cid:48) on the figure. The full field of view is8 (cid:48) in diameter, and the lighter stripe in the middle of the image is due to the gap between the two SALTICAM chips. CD Ind is marked on the figure, and noshells or shell fragments are visible in the image. Right:
The unsharp masked image of CD Ind shown here was processed using the same steps described inthe caption for Figure 1. Again, it does not show any shells or shell fragments.MNRAS000