Spectroscopic Follow-up on Potential Magnetic Cataclysmic Variables

The system 0026+24 was discovered as a transient in the CSS survey (Drake et al. 2009) under the identifier CSS 091026:002637+242916. Margon et al. (2014) used the high variability in its light curve obtained with the Palomar Transit Factory, along with follow-up spectra that showed strong Balmer and He ii emission lines to identify this system as containing a magnetic white dwarf. No period was determined for 0026+24. The ZTF light curve (ZTF17aaaehby) shows a range from 17.5 to 20.5 mag (Figure 3). The magnitude at the time of our eight spectra on 2018 December 4 was 18.6, so at a mid-state. The spectrum in Margon et al. (2014) compared to ours (Figure 1) shows a bluer continuum and stronger He ii compared to Hβ than ours, which may be due to a lower level of mass accretion at the time of our spectra. The radial velocities showed periods close to 2 hr for He ii, Hβ, and Hα. A period search using the ZTF light curve revealed an eclipse and refined the period to 2.04760(2) hr (122.9 minutes). Figure 3 shows the power spectrum and light curve folded on this period. The presence of two peaks in the light curve (phases 0.1 and 0.6) is indicative of two accretion poles in view during the orbit. With this fixed period, the final solutions for the velocity curves were obtained and are shown in Table 2 and Figure 2. The large (>400 km s⁻¹) values for the K semi-amplitudes are typical of many polars (see Thorstensen et al. 2020 for a compendium of K velocities for some new polars). The eight spectra shown in Figure 4 are phased according to the spectroscopic phasing, where phase 0 is the red to blue crossing of the velocities. The sequence shows the broad base and narrow component structures that are characteristic of polars, as well as changes in the continuum shape. However, better time-resolved spectra on a larger telescope and polarimetry will be needed to further characterize the system parameters.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

This object was identified as a CV in the CSS (CSS 071108:030945+295251; Drake et al. 2014). Thorstensen et al. (2017) claimed a period of 2.03 hr and the presence of an eclipse. However, no spectrum was shown. The light curves of this object show a range of 15–19, and our spectra were taken at a magnitude of 17, midway between the high and low states. The spectra have low signal-to-noise ratio (S/N), with Hα in emission, but the higher-order Balmer lines have faint emission surrounded by broad absorption and no evidence of helium lines (Figure 1). Our 83 minutes of observation time on 2018 December 4 covers about half an orbit, with the resulting velocity curve for Hα showing a K semi-amplitude of 210 km s⁻¹ (Figure 2, Table 2). There is no evidence for a magnetic nature from these data.

The American Association of Variable Star Observers (AAVSO) lists this object as MGAB-V330 with an online light curve by G. Murawski and their classification as an eclipsing polar. However, their light curve appears to show high and low states that are typical for a polar, as opposed to eclipses. A light curve from ZTF17aaaheoj is consistent with this interpretation of states (Figure 5). The range in magnitude is 16.5 at a high state and near 20.5 during a low. Gaia shows a mean g magnitude of 19.6 and a parallax of 2.33 ± 0.07 mas, indicating a distance of 429 ± 18 pc. The APO spectra obtained on 2019 April 2 occurred during a high state. One set of blue and red spectra is shown in Figure 1.

Zoom In Zoom Out Reset image size

The spectra are typical of polars in a high state of accretion, showing Balmer emission lines with narrow components as well as strong He ii lines. Fitting the centroid of the lines to a sine curve produced periods of 89–98 minutes and large K semi-amplitudes near 400 km s⁻¹.

In order to obtain a more precise measurement of the orbital period, we used TESS data (10.17909/wjjd-h183). The source was observed during sector 59 (from 2022 November 26 to 2022 December 23) at a full-frame cadence of 200 s. The power spectrum shows a single peak at 15.627 cycles day⁻¹ (92.1 minutes), which we presume to be the orbital frequency. Fixing the period at this value produces the velocity curve fits shown in Table 2 and Figure 2. The values of period and velocity amplitude are typical of polars.

In addition, broad cyclotron humps that vary in amplitude among the spectra appear near wavelengths of 4300 and 7100 Å (Figure 1). To obtain some estimate of the magnetic field, we used the typical relation between the nth harmonic of the cyclotron and the magnetic field B in megagauss (Ferrario et al. 2015):

$$\begin{eqnarray}&&{\lambda }_{n}=\displaystyle \frac{10710}{n}\,\left(\displaystyle \frac{100\text{MG}}{B}\right)\;\sin \;\theta \ {\rm{\mathring{\rm A} }},\end{eqnarray} \tag{ 1 }$$

where λ_n is in angstrom and θ is the viewing angle of the cyclotron beaming.

A field strength of 50 MG has the 5th harmonic at 4280 Å and the 3rd harmonic at 7134 Å (the 4th harmonic at 5350 Å is not covered in our spectral range), which is a reasonable fit to our data. Thus, 0548+53 is confirmed as a bona-fide polar.

This CV (referred to as SDSS J083751.00+383012.5 in the AAVSO Variable Star Index, VSX, catalog) was discovered in the SDSS survey (Szkody et al. 2005), and follow-up photometry, spectroscopy, and spectropolarimetry were reported in Schmidt et al. (2005). Those follow-up observations were all done during a low state that lasted for over 3 yr. Two cyclotron humps were present, indicating a field strength of 65 MG, and the presence of positive and negative polarization showed that both poles were visible during portions of the orbit, and the low amplitudes of variability indicated a low inclination for this system. The power spectrum of the two nights of photometry gave periods of either 3.18 or 3.65 hr, with similar significance, while a fit to a sine wave gave a slightly better fit for the 3.18 hr period.

Our spectra in 2021 were obtained during a high state, with prominent Balmer and He ii emission lines (Figure 1). The velocity fits are given in Table 2 and shown in Figure 2. The best fit for the lines was obtained with a period of 3.18 hr (190.8 minutes), confirming this as the correct period for this system. The fact that it does have high states of accretion shows that this is a regular polar system, not one of the low-accretion-rate polars. The small radial velocity amplitude for a polar could be due to a low inclination to the line-emission area.

This SDSS discovered object was proposed as a likely polar because of its very strong He ii emission (Szkody et al. 2004). The SDSS spectrum was obtained during a high state. Follow-up spectra obtained in 2015 June (Szkody et al. 2018) had a flux that was about 5 times fainter, with only narrow Balmer emission and little to no ionized helium, indicating a low state. Our five spectra in 2022 May spanned 92 minutes and were similar to the high-state SDSS spectrum.

We searched the ZTF light curve (ZTF18aditqga) for periodic signals. Using the Lomb–Scargle method, we did not find any significant period. However, after detrending the light curve and using a box least-squares method, we did find one significant period, 3.17 hr; see Figure 6. The light curve folded on this period shows a subtle but significant partial eclipse with a depth between 10% and 20%. Using this period, we attempted to fit the radial velocities with the 92 minutes of coverage from the five spectra. The best result was obtained for the Hβ line, with the result listed in Table 2 and shown in Figure 2. Given the limited spectral coverage, further data will be needed to determine the correct identification of system parameters. With the period between 3 and 4 hr, the relatively flat light curve outside the partial eclipse, and the lack of any signal indication of a spin period, it is likely that 1626+33 is a high-accretion-rate SW Sex star, rather than a polar or IP.

Zoom In Zoom Out Reset image size

As pointed out in Szkody et al. (2020), this system has not been studied in detail, although it was first identified as a CV by Appenzeller et al. (1998). The previous ZTF light curve shown in Szkody et al. (2020) and the recent one (Figure 9) reveal high and low states between 17.5 and 20.5, and spectra taken in 2019 May during a high state showed strong He ii and Balmer emission that changed shape over the course of the hour. Our further spectra obtained in 2021 (57 minutes at APO and 73 minutes at Pal) and in 2023 (148 minutes at APO) also during the high state were similar in appearance to the 2019 data. However, no consistent period could be extracted from the radial velocities alone as large changes in velocity were evident throughout the data sets.

TESS has observed this source extensively, including two separate ∼year-long intervals at the default FFI cadence. Using TESSCut, we downloaded FFI cutouts for 1631+69 in sectors 40, 41, and 47–60 (10.17909/bdjn-y925). The cadence between consecutive points in these light curves was usually 10 minutes, but in the final two sectors it was just 200 s. While TESS data from sectors 14–26 are also available, we ignored them because their 30 minutes cadence is too slow for an analysis of the fast variability in 1631+69.

Although the source was undetected in the data in sectors 56–58, a power spectrum of the remaining sectors reveals the likely orbital and white dwarf spin frequencies as well as their harmonics and sidebands (Figure 7). While the frequency identifications are ambiguous, we believe that the most probable orbital frequency is 15.4682 cycles day⁻¹ (93.1 minutes) based on several circumstantial arguments and reasonable assumptions:

• 1.  
  We assume that the fundamental orbital and spin frequencies are both above the noise in the TESS power spectrum.
• 2.  
  The orbital period is presumed to be above the ∼78 minutes (∼18.5 cycles day⁻¹) period minimum for hydrogen-rich systems.
• 3.  
  We disallow identifications in which a fundamental frequency has a subharmonic. For example, the signal at 20.0 cycles day⁻¹ in Figure 7 cannot be the fundamental spin, orbital, or beat frequency because it is the second harmonic of a frequency at 10.0 cycles day⁻¹.
• 4.  
  We presume that the white dwarf spin frequency is above the binary orbital frequency (as is observed in all well-studied magnetic CVs except for the asynchronous polar V1432 Aql).
• 5.  
  Finally, we expect the frequencies in the power spectrum to generally accord with theoretical predictions and observational results of other IPs.

Zoom In Zoom Out Reset image size

While these considerations significantly narrow down the possible frequency identifications, the orbital frequency could be either 15.4682 cycles day⁻¹ or 10.0011 cycles day⁻¹. Fortunately, the absolute Gaia G magnitude of 1631+69 and its G_BP − G_RP colors can be used to constrain the orbital period of the system, and it is on this basis that we believe that 15.4682 cycles day⁻¹ is the orbital frequency. The Gaia early Data Release 3 distance (${510}_{-110}^{+140}$ pc; Bailer-Jones et al. 2021), in conjunction with the negligible line-of-sight extinction in the Green et al. (2019) reddening maps and the mean Gaia G magnitude of 20.2, results in an absolute Gaia G magnitude of ${11.6}_{-0.5}^{+0.6}$. Abrahams et al. (2022) found that the G_BP − G_RP color and the absolute G magnitude correlate with the orbital period of a CV, and, based on their Figure 1, G = 11.6 and G_BP − G_RP = 0.18 are strongly suggestive of an orbital period well below the period gap. The peak at 15.4682 cycles day⁻¹ is the only signal that satisfies both this constraint and our previous requirement that the system have an orbital period above the period minimum.

For an orbital frequency of Ω = 15.4682 cycles day⁻¹, the only plausible identification of the spin frequency is ω = 25.4693 cycles day⁻¹ (56.5 minutes). Assuming that ω > Ω, any other identification of ω would result in subharmonics of either ω or ω − Ω. The resulting spin-to-orbit ratio of 0.607 is anomalously large for an IP, making this system a new member of the rapidly growing class of slowly rotating IPs; see Littlefield et al. (2023, their Figure 5).

Although the McDonald 2.1 m photometry (Figure 8) lacks the year-long baseline of the TESS data, it has superior time resolution and S/N. The 2(ω − Ω), ω, and Ω frequencies in the TESS power spectrum were observed in the McDonald 2.1 m data during 2020 July–September, although the power spectrum varied from night to night. The July 24 data showed the 2(ω − Ω) at 72 minutes, while the August 18 data instead showed two significant periods at 54 and 93 minutes, corresponding with ω and Ω, respectively. Using the precise period from the TESS observations, we are able to derive a long-term ephemeris for the proposed orbital period. Further refinement of the period was accomplished using the McDonald 2.1 m observations, and a zero-point was determined from the dips seen in the light curve in Figure 8, which shows the six nights of McDonald data folded on the adopted (tentative) orbital ephemeris:

$$\begin{eqnarray*}&&{T}_{0}=\mathrm{BJD}\ 2459054.718960\,(2)+0.064648680\,(5)\,E.\end{eqnarray*}$$

Zoom In Zoom Out Reset image size

Figure 9 shows the ZTF light curve, power spectrum, and orbital phased data. Because the sampling rate of ZTF is very different than TESS, the power spectrum shows different alias frequencies, but frequencies previously identified as the orbital frequency and spin frequency are both present in the ZTF power spectrum.

Zoom In Zoom Out Reset image size

The McDonald (Figure 8), ZTF (Figure 9), and TESS (Figure 10) data all show a dip in the light curve folded at the orbital frequency Ω, which we identify as a grazing eclipse of the accretion flow. By phase-folding the TESS data, we can disentangle the orbital and beat modulations and glean insight into the highly variable depth of the eclipse in the McDonald data. Figure 10 presents a two-dimensional light curve in which the orbital profile is shown as a function of the system's spin–orbit beat phase. The deepest eclipses occur only at a narrow subset of beat phases.

Zoom In Zoom Out Reset image size

Returning to the spectra, even if we force the spectroscopic period to match either of the photometric periods (72 or 93.1), we still cannot produce consistent radial velocity solutions. The best fit occurred for Hα in the 2021 April data (listed in Table 2 and shown in Figure 2). The Hα and Hβ lines from the 2019 data showed comparable values of semi-amplitude (140 ± 43 and 122 ± 24 km s⁻¹, respectively) lending some support to a value near 100 km s⁻¹. However, the longer observation length in 2023 resulted only in a 2–3σ fit with low K amplitudes (43 ± 15 km s⁻¹). It is clear that the K amplitude is not large, even though there is a grazing eclipse. As the emission lines likely arise in the stream/magnetosphere region, it is not so surprising that an orbital modulation is difficult to measure accurately if the inclination is not favorable. The spectra on the night of 2023 May 24 were obtained simultaneously with the McDonald photometry, and included the sharp eclipse at cycle 15994 shown in Figure 8. The eclipse spectrum is fainter and has the reddest velocities for both Hα and Hβ. The spectrum obtained during the brightness peak just after the eclipse has the strongest lines with the continuum only slightly increasing. These results are consistent with an eclipse of a portion of the accretion stream by the secondary as it flows to the white dwarf.

The variability of this object was detected by both the ATLAS and ZTF sky surveys (Heinze et al. 2018; Ofek et al. 2020). Detailed analysis of the photometry by F. Romanov was published on 2021 February 23 in the AAVSO VSX (Watson et al. 2006) with the name Romanov V48 and later published in Kato & Romanov (2022). No spectra have been published. The photometric data show three periods, identified by Kato & Romanov (2022) as an orbital one at 147.3 minutes, a spin period at 60.62 minutes, and a beat period at 103 minutes. These periods indicated a classification as an IP with a large P_spin/P_orb ratio of 0.41. However, an IR excess suggested cyclotron emission and perhaps a polar origin, so the authors suggested this could be an object between an IP and a polar. The long-term ZTF18aawrcla light curve (Figure 11) shows the prominent periods previously identified.

Zoom In Zoom Out Reset image size

Our spectra from Lick and APO (Figure 1) show the strong He ii indicative of IPs and polars. However, as in 1631+69, the radial velocities of the emission lines showed large scatter (Figure 2), even with fixing the period at 147.3 minutes and the K amplitude of the best fits (Table 2) being not large, indicating a relatively low inclination.

TESS observed 2011+60 in sector 41 (2021 July 23–August 20) and sectors 55–57 (2022 August 5–October 29) at a 2 minutes cadence (10.17909/0yb0-m366). While it was also observed in several earlier sectors, the previous data were obtained at a 30 minutes cadence, so we exclude the older TESS observations from our analysis. A power spectrum of the 2 minutes data shows many periods that are combinations of the spin and orbit periods proposed by Kato & Romanov (2022; see Figure 12).

Zoom In Zoom Out Reset image size

We independently arrive at the same frequency identifications proposed by Kato & Romanov (2022). Using the Abrahams et al. (2022) method to predict the orbital period from the Gaia absolute G magnitude and G_BP − G_RP colors, we would expect the orbital period to be ∼3 hr, and only one signal in the observed power spectrum matches this prediction: the 9.77 cycles day⁻¹ (147.3 minutes) frequency identified by Kato & Romanov (2022) as the orbital frequency. Additionally, the Kato & Romanov (2022) identification of the spin period is significantly more reasonable than other choices; for example, if the 2ω − Ω signal in Figure 12 were actually the true spin frequency ω, the resulting beat frequency ω − Ω would have a subharmonic, which is implausible with a Lomb–Scargle spectrum. While we believe the Kato & Romanov (2022) frequency identifications to be correct, they nonetheless require confirmation of their origins, as with 1631+69.

Notably, the likely orbital period at 147.3 minutes places 2011+60 in the period gap; with the solitary exception of Paloma, all IPs with large P_spin/P_orb ratios are below the period gap (Figure 5 in Littlefield et al. 2023).

When the spin period is phased with the beat cycle (Figure 13), large changes in the spin profile are evident, as first noted by Kato & Romanov (2022) from their analysis of the ZTF data. The TESS data show this effect in more detail.

Zoom In Zoom Out Reset image size

While a disk-fed IP would show a more uniform spin modulation of the two accretion curtains, Figure 13 shows some evidence of the two poles (beat phases 0.88–0.04), whereas at other times (beat phases 0.08–0.16, 0.28–0.40), there are sharp peaks present in the pole visible at spin phase 0.3. This likely means that there is some input from the mass-transfer stream to the magnetosphere.

These data all confirm 2011+60 being a short-period member of the IP class.

Margon et al. (2014) identified 2313+16 as a potential magnetic CV based on its spectrum showing strong He ii emission and the PTF light curve that revealed a period of 1.36 hr (81.6 minutes). We were only able to obtain four spectra, but they cover 0.52 of its orbit so we were able to construct a radial velocity curve (Figure 2). The results shown in Table 2 reveal a high K semi-amplitude that is typical of polars, as well as large changes in the He ii strength during the 53 minutes of observations. This lends support to a classification of 2313+16 as a polar, while the skewed appearance of the light curve both in Margon et al. (2014) and in our ZTF folded light curve (Figure 14) is more typical of an IP. Our spectrum shown in Figure 1 is fairly similar to that shown in Margon et al. (2014), although the continuum shape is different, with less blue light in the APO spectrum. However, the Margon et al. (2014) spectrum was normalized so it is difficult to do a direct comparison, and the phase coverage may also be different.

Zoom In Zoom Out Reset image size

In analyzing the ZTF data, we clearly recovered the orbital period (Figure 14). We also detected a significant peak in the power spectrum at 33.3 day⁻¹ (43.19 minutes), which we originally thought could be a spin frequency. It is 1.9 times the orbital frequency and therefore not a harmonic of the orbital period. However, the TESS power spectrum of 2313+16, which was obtained 2022 September 1–30 from 200 s of cadence FFI data in sector 56 (10.17909/g2jp-6m34), shows only orbital harmonics, with no evidence of the 33.3 day⁻¹ signal (Figure 15). The most likely explanation for this is that the 33.3 day⁻¹ signal in the ZTF data is an alias of the second harmonic of the orbital frequency, as it is exactly 2 day⁻¹ below that harmonic.

Zoom In Zoom Out Reset image size

The folded TESS data (Figure 15) show a sharp dip that may be indicative of a partial eclipse, as well as a broader dip a quarter of a phase later that could be caused by a stream or the effects of a second pole. Polarimety would be very helpful to clarify the correct classification of this object.