1. Introduction

Dwarf novae are a sub-set of cataclysmic variables - binary systems hosting a white dwarf and a low-mass main sequence star. In these systems, the companion is transferring mass to the white dwarf via Roche-lobe overflow and an accretion disk is formed around it. Dwarf novae can undergo out-bursts, thought to occur due to thermal instabilities in the disk (see ^{Lasota 2001}, for details) resulting in a brightness increase of ≈6-100 times than its quiescence level (^{Osaki 1996}). While these dwarf novae outbursts can be very homogeonous from a single source, occasionally they exhibit ‘superout-bursts’. These are different from normal outbursts as they are brighter and can last 5-10 times longer (≈15-40 days compared to ≈2-20 days in normal outbursts, the exact values depend on the type of system; ^{Mauche et al. 2001}).

MASTER OT 075353.88+174907.6 (hereafter OT 0753) was reported as a optical transient on 2017 October 31 (^{Balanutsa et al. 2017}) by the MASTER-Net project (^{Kornilov et al. 2012}). This transient had a B magnitude of 18.6 during the discovery observation. The Sloan Digital Sky Survey (SDSS) had previously observed the source position of OT 0753 and did not detect any source. The sensitivity limit of the SDSS in the g’ band (which is the closest one to the Johnson B band from the MASTER-Net project) is ≈23 mag. This indicates (see also ^{Balanutsa et al. 2017}) that during the MASTER-Net detection OT 0753 was ≳4.4 mag brighter than during its known quiescent level. Furthermore, a NUV detection within 5 arcsec of this source position was found in the Galaxy Evolution Explorer mission (GALEX; ^{Martin et al. 2005}) database. This source was observed on 2006 February 7 using GALEX and during that observation it had a magnitude of ≈22.9 in the NUV (^{Balanutsa et al. 2017}). The source was not detected in the GALEX/FUV band. The GALEX/NUV detection and SDSS non-detection suggest that the source hosts a blue object when in quiescence.

To investigate the nature of this source, we carried out a multiwavelength campaign across the near-infrared, optical, UV, and X-ray bands during its 2017 outburst. We discuss the results of our observing campaign and show that this blue optical transient is likely a superoutburst of a dwarf nova which hosts a hot white dwarf that dominates the emission in quiescence.

2. Observations and data analysis

We requested follow-up observations of OT 0753 using the Neil Gehrels Swift Observatory (^{Gehrels et al. 2004}). We observed OT 0753 a total of 9 times using Swift (see Table 1 for a log of these observations). We obtained the first observation within a day of the initial report of the transient (see also ^{Parikh & Wijnands 2017}). This observation was carried out over all six optical and UV bands using the Ultraviolet and Optical Telescope (UVOT; ^{Roming et al. 2005}). Subsequent observations using the UVOT only observed the source in the three UV bands. Further UV and X-ray data of the source were obtained using Swift up to ≈1 month after the initial detection, until the source had approached its assumed quiescent level.

Fig. 1 The detection of OT 0753 (indicated by the dashed black circle) in outburst and its subsequent quiescence phase is shown. Top two panels: the RATIR images in the optical g’ band; bottom two panels: the Swift/UVOT images in the UV m2 band. The color figure can be viewed online. 

We analysed all the Swift/UVOT data using the Level 2 products. We added all the individual exposures (in the same band) of a given observation using uvotimsum. We selected a circular source extraction region having a radius of 5 arcsec. The background used was a circular region having a radius of 10 arcsec, placed on a source-free location on the CCD. The source magnitude and flux were extracted using the uvotsource tool. The results of this analysis is shown in Table 1.

The source was not detected in X-rays through-out our observing campaign. The upper limits on the source count rate (in the 0.5-10 keV energy range) during the individual observations taken with the X-ray Telescope (XRT; ^{Burrows et al. 2005}) observations were ≲(0.8-2.4)×10-3 counts s⁻¹ (depending on the exposure time of each observation; calculated using the 90% prescription by ^{Gehrels 1986}). To obtain the flux upper limits, we simulated X-ray spectrum using Xspec (^{Arnaud 1996}) employing an absorbed power-law model and assuming that a power-law index of 2 was representative of the spectrum (similar to what has been observed from cataclysmic variables that were detected in the X-rays; ^{Done & Osborne 1997}; ^{Balman 2015}). The equivalent hydrogen column density N_H was obtained using the HEASARC N_H tool³ and was found to be 3.8 × 10²⁰ cm⁻². Our obtained count rate upper limits corresponded to flux upper limits of ≲(4-9)×10-14 erg cm⁻² s⁻¹ in the 0.5-10 keV range. We stacked all the XRT observations using ximage to check whether the source could be detected if we used all available data. We found that the source was still not detected in ≈13.4 ksec of stacked XRT images and we obtained a stricter X-ray upper limit of ≲1.7×10-4 counts s⁻¹ (corresponding to a flux upper limit of ≲6×10-15 erg cm⁻² s⁻¹, again assuming a photon index of 2).

The optical and near-infrared follow-up was carried out by a ground-based campaign using the 1.5 m RATIR (Reionization and Transients InfraRed) telescope located at the Mexican Observatorio Astronómico Nacional (OAN) on the Sierra San Pedro Mártir (SPM) in Baja California (^{Watson et al. 2012}). This campaign lasted from 2017 November 12 to 2017 November 25 and included 12 nights of observations with exposure times ranging from ≈0.8-2.4 hours per night (see Table 1 for a log of the RATIR observations). On each night, the source was observed using the SDSS g’, i’, and z’ bands. We analysed the RATIR data using the PYTHON package PhotoPipe⁴. This pipeline automatically reduced all the data, carrying out the following processing steps: (1) calibrating the images using the bias, dark, and flat fields, (2) performing astrometry, and (3) co-adding the data and performing absolute photometry. For both the RATIR as well as the UVOT results, all errors are given for the 1σ confidence level and all magnitudes correspond to the AB system. The light curve constructed using these data is shown in Figure 2.

Fig. 2 The evolution of OT 0753 across the various optical/near-infrared (upper panel) and UV bands (lower panel) is shown. The upper panel also shows the detection of OT 0753 as reported by the Gaia Photometric Alert (blue-green squares) and the MASTER-Net project (blue diamond; ^{Balanutsa et al. 2017}). The SDSS detection limit is shown in the upper panel and the GALEX/NUV detection level (along with its error bar as indicated by the grey band) is shown in the lower panel. All magnitudes correspond to the AB system. The grey vertical dashed lines indicate the times for which we have constructed the SEDs shown in Figure 3. The color figure can be viewed online. 

3. Results

The first notice of the outburst of OT 0753 was published by the MASTER-Net Project on 2017 October 31 (in the B band, indicated by the blue diamond in the upper panel of Figure 2; ^{Balanutsa et al. 2017}). The rise to outburst of OT 0753 was not detected and the exact day on which the out-burst began is not known. We examined the Gaia alerts⁵ and found that it was detected by Gaia on 2017 October 28 (≈3 days before the reported detection from the MASTER-Net project; the source was observed twice using the Gaia G broadband filter, indicated by the blue-green squares in the upper panel of Figure 2). However, this information was not immediately publicly available and could only be retrospectively obtained. We examined the pre-outburst GALEX and SDSS data for this source. We assumed that the archival GALEX data (detected in the NUV and not detected in the FUV) and the non-detection by the SDSS are representative of our system in (pre-outburst) quiescence (see also ^{Balanutsa et al. 2017}). These data are summarized in Table 1. We estimated the upper limit of the source magnitude in the GALEX/FUV band using the known sensitivity versus exposure time relation⁶.

The RATIR and Swift/UVOT light curves are shown in Figure 2. The outburst displayed a slowly decaying phase that lasted ≳16 days from MJD 58054 to 58070. The first RATIR observation was only obtained near the end of this slow decay phase. OT 0753 was only detected in the i’ and z’ near-infrared bands using RATIR during this slowly decaying outburst phase. After MJD 58070 the optical magnitude rapidly decayed, by a factor ≈3 in ≈3-5 days, to values close to the SDSS upper limit determined when OT 0753 was (assumed to be) in quiescence. Around this time the UV magnitude also began to decrease but this decrease was slow compared to the optical decay. The UV magnitude only decreased by a factor of ≳1.5 over the same time (as compared to a drop by a factor ≈3 in the optical in ≈3-5 days). In spite of this decrease, during the next observations (at MJD 58075) the UV magnitude did not approach its assumed pre-outburst quiescent level and continued to decay slowly. The last observation of the source, carried out on MJD 58091, indicated an upper limit of ≲22.5 mag in the m2 band. This band is similar to the NUV band on board GALEX. Thus, our last UVOT observation of OT 0753 indicated an upper limit consistent with the known GALEX/NUV quiescent level (of ≈22.9±0.5 mag; see also Figure 2), suggesting that the source was (likely) back in quiescence.

We constructed several spectral energy distributions (SEDs; see Figure 3) of OT 0753 to understand the broadband evolution of the source during its outburst. The (quasi-)simultaneous SEDs were constructed at the times indicated by the vertical dotted grey lines in Figure 2. We calculated the spectral indices (Γ) corresponding to these SEDs using the relation

Fig. 3 The SED evolution of OT 0753 at various times is shown. The vertical dotted grey lines in Figure 2 indicate the times for which these SEDs were constructed. The assumed quiescent level using the pre-outburst GALEX and SDSS data is shown in orange. The color figure can be viewed online. 

The SEDs corresponding to the broadband coverage of the initial slow outburst decay (over MJD 58059 to 58070) are shown in Figure 3 (as black •, green ◆, and red ■, respectively; upper limits are always shown using ∇ in the appropriate colour). These SEDs fit well a blackbody model. The spectral indices corresponding to these SEDs were Γ ≈1.5- 2.3 and such values are consistent with the expected emission from a disk-like spectrum (^{Frank et al. 2002}).

Several observations during the subsequent rapid outburst decay stage (after MJD 58070) yielded only upper limits with very few detections across the various bands. The SED constructed for the observations around MJD 58075 is unconstrained with Γ≲4.1 (plotted as magneta ★ in Figure 3).

Further, we examined the SED shape around MJD 58080 and 58081 (near the end of the rapid optical and corresponding slow UV decay phase, as shown by the dark blue ✴ and light blue ◦, respectively, in Figure 3). Unfortunately, detections were only obtained in either the UV or the optical bands for each of these days. However, the source magnitude level at this stage of the outburst is not expected to vary much across time scales of day.⁷ Thus, we have a quasi-simultaneous detection in the optical and UV bands which suggests a very steep SED, with Γ≈3.7±0.7.

The discussion in § 4 shows that this very steep SED a few days after the end of its outburst might not be unrealistic for the quiescent state of OT 0753. Fitting a blackbody model to this SED indicated a temperature of ≳20,000 K. The actual temperature may also be indicative of significantly higher temperatures since for an SED corresponding to these high spectral indices our coverage over the optical and UV bands only probes the tail of the Rayleigh-Jeans distribution (Fλ∝λ-4). A further increase in this temperature does not significantly change the spectral index calculated over the wavelength range we probe.

If the pre-outburst flux is indeed representative of the quiescent flux level of the source, then the source flux must drop further after MJD 58080-58081. This is confirmed by the UV upper limit we obtained around MJD 58091 (see Table 1 and Figure 2) which indeed demonstrates that the source decreased further (at least in the UV).

We also constructed an SED at the assumed pre-outburst quiescent level, using the GALEX and SDSS data, as is shown in orange in Figure 3. The spectral index obtained for our assumed pre-outburst quiescent level is 5.3≳Γ≳2.5. The upper limit determined using the GALEX/FUV band is unconstraining. We fit this pre-outburst quiescent level SED with a blackbody model which results in a blackbody temperature ≳12,000 K.

4. Discussion

We report on the near-infrared, optical, and UV behavior of the transient source OT 0753 during its 2017 outburst and subsequent quiescence. The broadband spectral behavior of the source, the length of the outburst (≳19 days), and the increase in magnitude above its quiescent level (≳ 4.4 mag) indicates that this is likely a superoutburst of a dwarf nova.

^{Otulakowska-Hypka et al. (2016)} have presented the statistical properties of several dwarf nova by studying their outbursts. We use their study to learn more about OT 0753 based on its properties that we observe. We note that their study is based only on optical data and therefore we will only use properties determined from our optical observations to compare them to their results. The duration of the superout-burst of OT 0753 in the optical was ≳19 days. From this value we can infer (using the trend they observe of the duration of the superoutburst versus the amplitude of the superoutburst; see their Figure 16) an expected optical amplitude increase of ≳4 mag which is consistent with what we have observed for OT 0753. This supports the inference that OT 0753 is a superoutburst of a dwarf nova.

The 2017 superoutburst is the first reported out-burst from this system. Using the relationship between the amplitude of the superoutburst and the amplitude of the normal outburst presented by ^{Otulakowska-Hypka et al. (2016}, see their Figure 13), we find that the amplitude of the normal outburst is ≈1-1.5 magnitude fainter than the superoutburst. This indicates that normal outbursts of OT 0753 will only have a magnitude of ≈20 in the optical bands meaning that the peak amplitude of such normal out-bursts from this source may not be detectable by the various sky surveys. For example, the survey limit of the MASTER-Net project (based on its individual snapshots) is ≈20-21 mag and that of the Catalina Real Time Survey (^{Drake et al. 2009}), which also observes this part of the sky, is only ≈19-19.5 mag. This could explain why the superoutburst of this source was the first activity to be detected and reported and why any previous normal outbursts were missed.

The relation between the amplitude of the super-outburst and the recurrence time between two consecutive superoutbursts presented by ^{Otulakowska-Hypka et al. (2016}, see their Figure 9) predicts that superoutbursts in this source should recur every ≈230 days. It is unknown why the source has not been observed before its 2017 superoutburst. During a superoutburst, similar to the 2017 one we study, the source is above the survey limit of several all sky surveys only for ≲5 days. Since the observation cadence of the source location by the various sky surveys is relatively sparse and since the number of days on which the source would be detected is small, it is not very surprising that the source has not been previously detected during a superout-burst. Alternatively, it could also be that OT 0753 has a superoutburst recurrence time longer than the ≈230 days predicted using the relation presented by Otulakowska-Hypka et al. (2016). Such a longer recurrence time is observed for sources that have a lower mass transfer rate (e.g., WZ Sge; ^{Lasota et al. 1995}).