Long-term Periodicities of Cataclysmic Variables with Synoptic Surveys

TL;DR

This study analyzes long-term periodicities in 344 galactic cataclysmic variables using combined PTF and CRTS data, revealing rare long-term variability likely caused by disk precession, triple systems, or magnetic changes.

Contribution

It is the first systematic analysis combining PTF and CRTS data to identify long-term periodicities in a large sample of CVs, exploring their possible mechanisms.

Findings

10 CVs exhibit long-term periodic variability.

Long-term periods are linked to disk precession or hierarchical triples.

Different mechanisms likely dominate for different period ranges.

Abstract

A systematic study on the long-term periodicities of known Galactic cataclysmic variables (CVs) was conducted. Among 1580 known CVs, 344 sources were matched and extracted from the Palomar Transient Factory (PTF) data repository. The PTF light curves were combined with the Catalina Real-Time Transient Survey (CRTS) light curves and analyzed. Ten targets were found to exhibit long-term periodic variability, which is not frequently observed in the CV systems. These long-term variations are possibly caused by various mechanisms, such as the precession of the accretion disk, hierarchical triple star system, magnetic field change of the companion star, and other possible mechanisms. We discuss the possible mechanisms in this study. If the long-term period is less than several tens of days, the disk precession period scenario is favored. However, the hierarchical triple star system or the variations in magnetic field strengths are most likely the predominant mechanisms for longer periods.

Tables

Table 1. Table 1: Observation log of our LOT observations

Source	Nights	Cadence	Exposure	Filter	$N$
	(day)	(second)	(second)
UMa 01	5	210	200	R	403
CT Boo	4	210	200	R	140
Her 12	2	310	300	R	105
Note: Cadence is the planned observation interval between each of the each frames, and is possibly affected by weather conditions or the instrument. $N$ is the number of the observations.

Table 2. Table 2: Summary of CVs having long-term periodicities

CV Nameaafootnotemark:	RA (J2000)	Dec (J2000)	Alternative Namebbfootnotemark:	Typeccfootnotemark:	$P_{\mathrm{orb}}$ddfootnotemark:	${\sigma }_{P_{\mathrm{orb}}}$eefootnotemark:	$P_{\mathrm{long}}$fffootnotemark:	${\sigma }_{P_{\mathrm{long}}}$ggfootnotemark:	$p\left(P_{\mathrm{long}}\right)$hhfootnotemark:	Amp.iifootnotemark:	Tspanjjfootnotemark:	Nkkfootnotemark:	$N_{\mathrm{ob}}$ llfootnotemark:
	(hh:mm:ss)	($\pm$dd:mm:ss)			(min)	(min)	(day)	(day)		(mag)	(day)	(#)	(#)
BK Lyn	09:20:11.20	+33:56:42.3	2MASS J09201119+3356423	DN/NL	107.97mmfootnotemark:	0.07mmfootnotemark:	42.05	0.01	1.88e-20	0.65	3242.92	514	—
CrB 06	15:32:13.68	+37:01:04.9	2MASS J15321369+3701046	NL	—	—	273.98	2.17	1.18e-13	0.11	3443.86	400	1
CT Boo	14:08:20.91	+53:30:40.2	—	NL	230.72	1.78	7.91	1.05	4.11e-04	0.13	2893.00	296	—
					78.65	0.36
LU Cam	05:58:17.89	+67:53:46.0	2MASS J05581789+6753459	DN	215.95nnfootnotemark:	0.0005nnfootnotemark:	256.76	2.10	3.18e-03	0.62	2934.95	363	—
Her 12	15:50:37.28	+40:54:40.0	SDSS J155037.27+405440.0	CV	75.62	0.008	326.81	1.32	4.97e-07	0.18	3033.88	429	1
					173.65	1.57
QZ Ser	15:56:54.47	+21:07:19.0	SDSS J155654.47+210719.0	DN	119.75oofootnotemark:	0.002oofootnotemark:	277.72	8.76	9.18e-07	0.09	3308.01	510	2
UMa 01	09:19:35.70	+50:28:26.2	2MASS J09193569+5028261	CV	404.10	0.30	246.84	0.81	8.01e-20	0.44	3251.05	601	1
V825 Her	17:18:36.99	+41:15:51.2	2MASS J17183699+4115511	NL	296.64ppfootnotemark:	2.88ppfootnotemark:	515.55	1.85	3.28e-16	0.16	3069.71	510	—
							9.24	0.05	4.51e-16	0.18
V1007 Her	17:24:06.32	+41:14:10.1	1RXS J172405.7+411402	AM	119.93qqfootnotemark:	0.0001qqfootnotemark:	170.59	0.12	2.13e-18	0.90	3069.71	507	—
VW CrB	16:00:03.71	+33:11:13.9	USNO-B1.0 1231-00276740	DN	—	—	142.60	0.24	1.92e-21	0.87	3271.19	351	$>$10

Table 3. Table 3: Periodicities in short-cadence observations

Source	$P_{\mathrm{opt}}$	${\sigma }_{P_{\mathrm{opt}}}$	$P_{\mathrm{peak}}$	$p$-value	N
	(min)	(min)
UMa 01	404.10	0.30	36.54	3.74e-18	403
CT Boo	230.72	1.78	9.80	2.99e-05	140
	78.65	0.36	9.49	4.29e-05	140
Her 12	173.65	1.57	26.75	1.00e-16	105
	75.62	0.008	19.22	6.02e-11	105
Note: $P_{\mathrm{opt}}$: optical periods, ${\sigma }_{P_{\mathrm{opt}}}$: errors in periods by Monte Carlo simulations, $P_{\mathrm{peak}}$: peak power in the power spectrum.

Table 4. Table 4: Long-term periodicities in CVs caused by superhumps

Name	$P_{\mathrm{orb}}$	$P_{\mathrm{sh}}$	$P_{\mathrm{long}}$	Type(sh)	Type	Reference
	(hr)	(hr)	(day)
AH Men	2.95	3.05	3.71	$+$	Nova	Patterson (1995)
DV UMa	2.06	2.14	2	$+$	DN	Vanmunster (2006)
CC Scl	1.38	1.44	1.6†	$+$	Nova	Woudt et al. (2012)
LT Eri	4.08	3.98	5.3	$-$	CV	Ak et al. (2005a)
LQ Peg	3.22	3.42	2.37	$+$	Nova	Rude & Ringwald (2012)
MV Lyr	3.19	3.31	3.6	$+$	Nova	Skillman et al. (1995)
NSV 1907	6.63	6.22	4.21	$-$	NL	Hümmerich et al. (2017)
PX And	3.41	3.51	4.43	$+$	CV	Thomas et al. (2010)
RR Pic	3.48	3.78	1.79	$+$	Nova	Schmidtobreick et al. (2008)
TT Ari	3.30	3.57	1.82	$+$	Nova	Stanishev et al. (2001)
TV Col	5.50	5.20	3.93	$-$	DQ Her	Augusteijn et al. (1994)
UX UMa	4.72	4.48	3.68	$-$	NL	de Miguel et al. (2016)
V1193 Ori	3.43	3.26	2.98	$-$	Nova	Ak et al. (2005b)
V2051 Oph	1.50	1.54	52.5†	$+$	DN	Vrielmann & Offutt (2003)
V603 Aql	3.32	3.47	3.15†	$+$	Nova	Suleimanov et al. (2004)
WZ Sge	1.33	1.37	5.7	$+$	DN	Patterson et al. (2002)
Note: $P_{\mathrm{orb}}$, $P_{\mathrm{sh}}$, $P_{\mathrm{long}}$ are orbital, superhump and long-term disk precession periods, respectively. Type(sh) is the type of superhump: “$+$” for positive superhumps, “$-$” for negative superhumps.
† average value of multiple periodicities

Equations

$$p\equiv \mathrm{Prob}(P>P_i)={\left(1-\frac{2P_i}{N-1}\right)}^{\frac{N-3}{2}}$$

$$M=c_1\sin (2\pi \varphi +c_2)+c_3\sin (4\pi \varphi +c_4)+c_5$$

$$ϵ\equiv \frac{P_{\mathrm{sh}}-P_{\mathrm{orb}}}{P_{\mathrm{orb}}}$$

$$P_{\mathrm{long}}=P_{1,2}\left(\frac{a_3}{a_{1,2}}\right)\frac{m_1+m_2}{m_3},$$

$$P_{\mathrm{long}}\approx \alpha \frac{P_3^2}{P_{1,2}}\frac{m_1+m_2+m_3}{m_3}(1-e_{\mathrm{out}}{)}^{3/2},$$

$$P_{\mathrm{long}}=K\frac{P_3^2}{P_{1,2}},$$

Full text

cataclysmic variables, binaries: close, surveys, catalogs, methods: data analysis, observational

1 Introduction

A cataclysmic variable (CV) is an accreting binary system composed of a white dwarf (WD) as primary and a low mass companion. In general, when the companion fills its Roche lobe, the mass flow stream passing through the inner Lagrangian point ($L_{1}$) generates an accretion disk around a non-magnetic WD. On the other hand, only a truncated disk can be formed in an intermediate polar (DQ Her type), a subclass of CVs with a highly magnetic WD. The WD in the polar (AM Her type) system has an even higher magnetic field that can prevent the formation of the accretion disk. The variability on different time scales of CV systems is caused by different mechanisms. The orbital periods of CVs typically range from 70 min to 24 h, which is strictly related to the binary separation and mass ratio. A census on the orbital period distribution of CVs reveals a period gap of approximately $2-3$ h (e.g., Warner, 1995), which is explained by the evolution scenarios of CVs.

The variation with a time scale longer than a day is typically called the superorbital or long-term variation in CV systems. Long-term variation has been detected in only a few CVs, and this has been neglected in subsequent research. Kafka & Honeycutt (2004) studied 100 CVs using the structure function to characterize the time scales of the long-term variabilities. However, no further results and the implications behind the long-term variabilities were addressed in this study. Various types of mechanisms were proposed to explain the long-term variations of CVs. For example, Thomas et al. (2010) discovered a long-term modulation with a period of $4.43\pm 0.05$ d in cataclysmic variable PX And through eclipse analysis, which was considered to be the disk precession period that triggers the negative superhump in this system. On the other hand, from the analysis of eclipse time variations in eclipsing binary DP Leo, a third body with an elliptical orbit and a period of $P=2.8\pm 2.0$ yrs was found by Beuermann et al. (2011). Honeycutt et al. (2014) discovered a period of $\sim$ 25 d oscillations, regarded as the result of accretion disk instability in V794 Aql during its small outbursts. Warner (1988) proposed that the cyclical variations of the orbital periods (on the time scale of years to decades) for some CVs are related to their quiescent magnitudes and outburst intervals, and the variations were inferred as the effect of the solar-type magnetic cycle of the companion. Kalomeni (2012) discovered several polars exhibiting long-term variability with a time scale of hundreds of days, likely caused by the modulation of the mass-transfer rate owing to the magnetic cycles in the companion stars.

Previous studies on the long-term periodicities of CVs are sporadic. With the help of recent large synoptic surveys, we are able to search for and further characterize the long-term variations of the CVs systematically. In Section 2, we introduce the synoptic survey projects we utilized, and the corresponding intensive observations made using the Lulin One-Meter-Telescope (LOT). Descriptions of our analysis method for the long-term periodicity of the sources are presented in Section 3. The possible mechanisms driving the long-term variability and further implications are presented in Section 4. In Section 5, we summarize the long-term periodicities from our results and discuss some of the previous studies related to the sources.

2 CV Catalog, Synoptic Surveys and Observations

The CVs selected for this study are from the catalog by Downes et al. (2006). The data used for this study are from two surveys: the Palomar Transient Factory (PTF) and the Catalina Real-Time Transient Survey (CRTS). In addition, the LOT, a small telescope capable of intensive observations, was utilized for finding the orbital periods of the targets that were unknown before this study.

2.1 CV Catalog for Source Matching

The CV catalog produced by Downes et al. (2006) (hereafter Downes’ catalog) is taken as a reference catalog for our study (see Downes et al., 2001, for the catalog description). The latest version of the catalog contains 1830 sources, including 1580 CVs and 250 non-CVs. The catalog used the General Catalogue of Variable Stars (GCVS) name as the source identifier. However, some of the sources have no GCVS names, and so the constellation name of the sources was adopted as the identifier. In this study, if multiple sources without GCVS names are in the same constellation, then the constellation name with distinct numbers were adopted as the project name of the source (e.g., UMa 01 for one of the CVs in the constellation Ursa Major). We used the Downes’ catalog for our matching process and then retrieved the light curves of the matches.

2.2 Palomar Transient Factory

The PTF (Law et al., 2009; Rau et al., 2009) project began observing in 2009.111http://www.ptf.caltech.edu/ The Samuel Oschin Telescope, a 48-in Schmidt telescope with Mould R, SDSS g′, and several H-$\alpha$ filters was adopted for the survey. The PTF and its successor intermediate PTF (iPTF) projects were accomplished in March 2017. A 7.9 square degree field of view was achieved with the camera configuration of PTF and iPTF. The next generation of the PTF project, called the Zwicky Transient Facility (ZTF), with its many upgrades in software and hardware, will be operated in mid-2017.

The software pipelines were readied for the real-time discoveries of transients (Masci et al., 2017) and for the generation of source detections (IPAC pipeline, Laher et al., 2014). The IPAC pipeline was developed by the Infrared Processing and Analysis Center (IPAC)222http://www.ipac.caltech.edu/, which reduced the images and generated the detection catalogs on a frame basis. The detection catalogs are stored in the IRSA archive.333http://irsa.ipac.caltech.edu We extracted the data from the local copies of the full photometric catalogs in IPAC. Metadata tables with information from the catalog headers were created for accessing the photometric data quickly. The light curve of a specific source could be retrieved out of the total of $\sim 24$ Terabytes within 2 min via our data retrieval pipeline.444The IPAC web interface with faster data retrieval process is currently online. Figure 1 shows a flowchart of the data-retrieval pipeline. A large area of the sky has been observed by the PTF project. The observation numbers in the density map of sky covered by PTF/iPTF are shown in Figure 2.

2.3 Catalina Real-Time Transient Survey

The CRTS project is conducted by analyzing data from the Catalina Sky Survey (CSS), which is originally designed for the study of asteroids. The CRTS team made use of the photometric data for studying the transient sky (Drake et al., 2009; Mahabal et al., 2011; Djorgovski et al., 2012).555http://crts.caltech.edu The data was gathered using three telescopes in Northern and Southern Hemispheres, including the Catalina Sky Survey (CSS, 0.7m), the Mt. Lemmon Survey (MLS, 1.5m), and the Sliding Springs Survey (SSS, 0.5m). No filter was adopted for the survey to maximize the discovery of asteroids. The CRTS data is available to the public through the Catalina Surveys Data Release 2 (CSDR2) website.666http://nesssi.cacr.caltech.edu/DataRelease/ The calibrated light curves can be accessed by users through the interface.

2.4 Lulin One-Meter Telescope

The orbital periods of the CVs are essential to the discussion on the mechanism of their long-term periodicities. For our sources of interest, only a portion of them have known orbital periods, as presented in Downes et al. (2006) and references therein. To investigate the targets with unknown orbital periods, we used the Lulin One-Meter Telescope, located in central Taiwan, to perform short-cadence observations. The LOT is a telescope with a field-of-view (FOV) of $11^{\prime}\times 11^{\prime}$. The limiting magnitudes with 5 min exposures are approximately $20-20.5$ mag in V and R bands. We used the R-band filter for our short-cadence observations. The 3 sources we observed with LOT are presented in Table 2.4.