Optical Time-series Photometry of the Symbiotic Nova V1835 Aquilae

Symbiotic stars (hereafter "SySt") are a rare subclass of cataclysmic variables consisting of a wide interacting binary containing a cool giant star and a very hot star surrounded by gas that is excited by the latter's short-wavelength radiation. In optical spectra, this produces the unusual combination of absorption features characteristic of cool star overlain by emission features from the hot gas (Mikołajewska 2003a; Mikołajewska et al. 2015), an unexpected combination that prompted the term "symbiotic." In many cases, the hot star is a degenerate object, a white dwarf or neutron star, while the cool star is a red giant or supergiant that may be tidally distorted by its proximity to the hot star. The cool star may emit a stellar wind, some of which interacts with the hot star to generate the excited circumstellar envelope and which accretes onto its surface. This complex arrangement can produce several types of brightness variations which may include ellipsoidal behavior, a "reflection effect" due to orbital motion, and irregular variations related to accretion. In some cases, the cool star pulsates, providing a Mira-like element to the light curve (Mikołajewska 2003a; Angeloni et al. 2014). Different stars exhibit different degrees of these behaviors, and are collected in the inhomogeneous "Z Andromedae" (ZAND) class of objects (Samus & Durlevich 2009). The AAVSO Variable Star Index (VSX) currently lists 157 stars with this classification. ⁵

An even rarer, related group is the NC class of slow novae (Samus & Durlevich 2009). The VSX lists only six such stars. Outbursts of up to 10 mag are characterized by a slow rise to peak light, which is sustained for about a decade and followed by a very slow decline. The spectra of some members can resemble those of SySt, and long-period light variations like those in SySt have been observed during quiescence. Mikołajewska (2010) provides an excellent description of these symbiotic novae.

A recent addition to the short list of slow, symbiotic novae is V1835 Aquilae. Earlier identified as NSV 11749 (Kazarovets et al. 2015), the star was first discovered to be variable by Luyten (1937). Williams (2005) derived photometry from plates in the Harvard collection showing an outburst circa 1903 that reached a photographic magnitude of 12.5 then slowly faded below the plate limit of m_ptg ≈ 15 mag by 1912. It remained below this limit through the end of the plate sequence in 1988, though it had been detected on a few deeper plates at ∼17 mag. Based on this outburst, Williams suggested that V1835 Aql is either a slow nova or a FU Orionis pre-main sequence star. Bond & Kasliwal (2012) presented optical and infrared spectra of V1835 Aql showing Balmer and helium emission lines on a continuous spectrum of a M1-M2 giant. Their spectrum also showed a broad emission line at 6825 Å caused by light from quintuply-ionized oxygen (O VI) at 1032 Å, Raman-scattering off of cold, neutral hydrogen, an effect only seen in SySt. Bond & Kasliwal (2012) provided an excellent discussion of the preceding interpretations of V1835 Aql and gave convincing evidence that this intriguing object is a symbiotic star.

V1835 Aql was detected independently by a survey for H-alpha emission (Kohoutek & Wehmeyer 1997) and the ChaMPlane project to detect cataclysmic variables (Grindlay et al. 2005; Rogel et al. 2006). Wehrung et al. (2013) monitored V1835 Aql photometrically in the VI passbands for six months in 2012 and saw slow variations that supported the assessment of Bond & Kasliwal (2012) that V1835 Aql is a symbiotic nova. However, the duration of that time-series was too short to determine the periodic nature of the star. Wehrung et al. (2013) also provided a spectrum of the star (which did not cover the SySt diagnostic line at 6825 Å) and argued for a reddening value of E(B − V) = 0.67 ± 0.10 mag. In addition, they found nine additional variable stars located nearby, many of which are now listed in the VSX under the name "[WLR2013] #" where the number corresponds to the identification shown in Table 2 of Wehrung et al. (2013).

In this paper, we report continued time-series CCD photometry of V1835 Aql aimed at clarifying the nature of this object in quiescence and determining its orbital period. We also update the photometric properties of the nearby variable stars and detect new ones. In Section 2 we describe the new observations, and outline the stellar photometry in Section 3. We analyze the light curve of V1835 Aql in Section 4, and discuss the light curves of the other variable stars in Section 5. We summarize our results in Section 6 and suggest future observations to further clarify the nature of V1835 Aql.

We obtained images of V1835 Aql between 2012 and 2017 using the Panchromatic Robotic Optical Monitoring and Polarimetry Telescopes (PROMPT) located on Cerro Tololo in Chile (Reichart et al. 2005). During this time, we used a variety of telescopes and CCD cameras, as summarized in Table 1, in order to maintain a regular monitoring cadence as the availability of the PROMPT equipment shifted over time; a secondary goal was to minimize equipment changes in order to maximize the internal consistency of our data set. Throughout this time, images were acquired using the Johnson V filter and the Cousins I filter. ⁶ Additional images were obtained contemporaneously with the VI images in the Johnson B filter between 2013 March 09 and 2014 September 26 (on a total of 51 nights), and in the Cousins R filter from 2012 August 5 to 2013 November 5 (40 nights). On most nights, three images were obtained in each filter, with the telescope position offset by a few pixels between each exposure. Observing nights were separated by 7–14 days in order to maintain a slow, regular cadence suited to the long periods of the variable stars found by Wehrung et al. (2013). All images were processed using the standard methods of bias, dark-current, and flat-field correction.

When combined with the early observations made with the 0.5 m telescope at Bowling Green State University (BGSU; see Table 1) described in Wehrung et al. (2013), ⁷ we have photometry from a total of 1190 images taken on 157 independent nights spanning the Julian dates of 2,456,064 to 2,457,933 days (5.1 yr). The median seeing of the V-band images was 25 full-width at half-maximum, and 27 for the I-band images, though in all filter sets the seeing varied from 16 to 37 at the 10% and 90% points of the seeing distribution.

We analyzed all the PROMPT images following the methods described in Wehrung et al. (2013). We obtained a list of star positions and instrumental magnitudes for each image using the DAOPHOT II point-spread function fitting software (Stetson 1987). These lists were merged using DAOMASTER and the instrumental magnitudes for each image were refined using ALLFRAME (Stetson 1994) utilizing a star list based on the best quality images. Using a uniform star list helps to mitigate the effect of blending and merging of stellar profiles resulting from the range of seeing conditions present in our images. We separated the images into sets based on the telescope, camera and filter used (see Table 1) and processed them separately with DAOMASTER and ALLFRAME. In doing so, we found that during the "P5AU" set, and again during the "P1AU" set, there was an abrupt rotation in the images (perhaps due to removal of the camera for maintenance) that complicated the DAOMASTER/ALLFRAME process; we split the images into separate pre- and post-rotation sets and found that the resulting DAOMASTER/ALLFRAME analysis ran well.

For each telescope-camera-filter set, we used the instrumental and standard magnitudes (m_i and m_s respectively) of fifteen in-field secondary standard stars established in Wehrung et al. (2013) to obtain the coefficients c₀ and c₁ in the transformation equation

$$\begin{eqnarray}&&{m}_{i}-{m}_{s}={c}_{0}+{c}_{1}\,{(V-I)}_{s},\end{eqnarray} \tag{ 1 }$$

where (V − I)_s is the standard color. The rms scatter around these least-squares fits were typically about 0.026 mag in V, about 0.044 mag in I, and 0.019 mag in R. For B, we did not have standard magnitudes from Wehrung et al. (2013), so we downloaded data for six uncrowded comparison stars from the APASS database (Henden & Munari 2014; Henden et al. 2015). ⁸ Though the formal uncertainty in the each APASS B magnitude was relatively large, ∼0.06 mag, the rms scatter around the B fit was only 0.031 mag, indicating a reliable calibration.

We then applied the coefficients to every star in the set to obtain the average standard magnitude of each star in each passband. The resulting color–magnitude diagram for the P5AA set is shown in Figure 1. We estimate the overall zero-point uncertainty in the photometric calibrations to be approximately 0.02 mag in V and I, and about 0.04 mag in R and B. The lower resolution of the APASS images and consequent potential for blending of faint neighbor stars with each comparison star leads to a concern that a systematic error in our B magnitude transformation could be present.

Zoom In Zoom Out Reset image size

To detect variable stars, we used the Welch & Stetson (1993) method on our paired V and I magnitudes from the 2.3 yr P5AA data set to search for correlated brightness changes. We found 17 variable stars, ten of which had been identified by Wehrung et al. (2013), including V1835 Aql. We also applied the Welch & Stetson (1993) analysis on the P1AU data set; though the data span a much shorter interval of 0.4 yr, the much larger field of view could contain many variable stars in addition to the three originally detected by Wehrung et al. (2013) from the BGSU data set (#85, #132 and #419). We again used the Welch & Stetson (1993) variability index to identify variable candidates in the P1AU set, but adjusted our selection criteria from those used in the P5AA search to focus on stars that show slow, correlated variations in V and I in their time-magnitude plots, as well as red mean colors. No doubt this has biased our search toward finding long period variable stars in the P1AU fields, within which we detected six new variable stars.

The names and equatorial coordinates of these 23 variables, along with their J and K_s photometry from the Two Micron All-Sky Survey's Point Source Catalog (Skrutskie et al. 2006), are listed in Table 2. The uncertainty in the J and K magnitudes for most stars is between 0.02 and 0.04 mag, the exception being the bright star #1 with J_err = 0.22 and K_err = 0.29 mag. The 2MASS photometry confirms that most of the new variable stars are quite red.

Table 2. Variable Star Coordinates and Literature Data

ID#	R.A. (J2000)	Decl. (J2000)	J	Ks	dkpc	Comment
417	19:07:42.4	+00:02:51	10.794	9.351	${31}_{-14}^{+147}$	V1835 Aql, dphot = 8–11 kpc
1	19:07:44.3	+00:07:10	4.916	3.308	${0.88}_{-0.04}^{+0.05}$	ASAS 190744+0007.1
26	19:07:55.1	+00:05:30	7.939	6.489	${3.9}_{-0.6}^{+0.9}$	⋯
27	19:07:37.5	+00:06:09	6.159	4.301	${3.2}_{-0.9}^{+1.9}$	IRAS 19050+0001
62	19:07:42.4	+00:02:22	8.397	6.920	${5.7}_{-1.2}^{+2.1}$	⋯
85	19:07:30.2	–00:02:51	8.470	6.965	${7.9}_{-2.2}^{+5.2}$	BGSU+P1 fields only
132	19:07:10.4	–00:01:55	8.379	6.841	⋯	BGSU+P1 fields only
244	19:07:25.1	+00:00:59	9.774	8.302	${17}_{-6}^{+21}$	⋯
255	19:07:48.5	+00:05:32	9.964	8.460	${24}_{-11}^{+106}$	⋯
419	19:07:17.2	+00:00:40	10.302	8.735	⋯	BGSU+P1 fields only
991	19:07:48.7	–00:01:07	10.689	9.688	${2.78}_{-0.18}^{+0.20}$	new
992	19:07:45.0	+00:04:05	9.578	8.006	⋯	new
993	19:07:55.3	–00:01:19	11.698	11.346	${2.22}_{-0.10}^{+0.10}$	new
995	19:07:26.7	+00:03:54	7.313	5.968	${2.32}_{-0.17}^{+0.19}$	new
996	19:07:42.8	+00:04:41	9.250	7.728	${7.3}_{-1.5}^{+2.7}$	new
987	19:07:22.5	+00:06:54	14.211	13.714	${1.49}_{-0.09}^{+0.10}$	new
988	19:07:23.3	+00:07:17	10.236	8.736	${15}_{-5}^{+17}$	new
105	19:08:13.7	+00:00:19	9.138	7.670	${5.8}_{-0.9}^{+1.4}$	new P1+BG
154	19:08:06.5	–00:03:13	8.637	7.199	${4.4}_{-1.0}^{+1.8}$	new P1+BG
184	19:07:11.2	+00:02:27	9.565	8.122	${16}_{-5}^{+14}$	new P1+BG
270	19:07:26.8	–00:03:47	9.947	8.565	${6.2}_{-0.8}^{+1.1}$	new P1+BG
358	19:07:36.8	+00:14:03	8.963	7.476	${9}_{-3}^{+13}$	new P1+BG
451	19:07:43.2	+00:08:24	7.828	6.133	${10}_{-5}^{+80}$	new P1+BG

Download table as:  ASCIITypeset image

Also shown in Table 2 is the distance to each star in kiloparsecs, computed from the trigonometric parallax listed in the early third data release "EDR3" (Gaia Collaboration 2021) of the Gaia space astrometry mission (Gaia Collaboration 2016). Cases in which the parallax error was more than half the parallax value are of little direct use, and are excluded from the table. The exception is V1835 Aql, for which we find a photometric distance of 8–11 kpc as described at then end of Section 4.

For each of these 23 variable stars plus the 15 comparison stars discussed above, we extracted the instrumental time-series photometry from ALLFRAME, and combined them differentially to obtain standard BVRI magnitudes. For a particular variable star on a specific image, we calculated

$$\begin{eqnarray}&&{m}_{s,v}={m}_{s,c}+{m}_{i,v}-{m}_{i,c}-{c}_{1}\,[{(V-I)}_{s,v}-{(V-I)}_{s,c}],\end{eqnarray} \tag{ 2 }$$

where the v and c subscripts refer to the variable and comparison star, respectively. We repeated this calculation for each of the N_c comparison stars visible on that image to arrive at N_c separate estimates of the variable star's magnitude, then took the median of the N_c estimates as the best estimate. We also computed the standard deviation of the mean (σ_m) from the N_c estimates as the best estimate of the uncertainty in the variable star's standard magnitude. Another estimate of this uncertainty, was produced by propagating the photometric errors from ALLFRAME and the uncertainty of the comparison stars' standard magnitudes through Equation (2). These values are listed in Table 3, along with the heliocentric Julian date (HJD) of the observation, the FWHM seeing (in arcsec), the airmass at the time of observation, and an integer on a 1–5 scale describing the visual quality of stellar profiles on the image (Q = 1 indicates round star with a radial profile plot having little scatter; Q = 5 marks a very elongated profile due to tracking errors and/or an out of focus stellar profile, both of which have large scatter in their radial profile plots). The photometry set is also indicated.

Table 3. Time-Series Photometry of the Variable Stars

Star	HJD	Mag		σm	Nc	Filter a	FWHM	Airmass	Qb	Set
417	2456064.7500	16.206	0.075	0.005	15	1	3.3	1.88	2	BGSU
417	2456064.7510	12.814	0.017	0.010	15	2	3.3	1.88	2	BGSU
417	2456064.7549	16.043	0.065	0.005	15	1	3.4	1.83	2	BGSU
417	2456064.7559	12.823	0.013	0.009	15	2	3.4	1.83	2	BGSU
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯
001	2456149.5079	11.705	0.020	0.021	6	1	4.4	1.43	3	P5AA
001	2456149.5476	9.698	0.011	0.011	5	3	3.8	1.24	3	P5AA
001	2456149.5480	7.304	0.049	0.030	5	2	1.9	1.24	2	P5AA
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯
451	2457700.5244	11.563	0.038	0.012	15	2	3.1	2.22	2	P1AU
451	2457700.5247	11.551	0.033	0.013	15	2	3.0	2.23	2	P1AU

Notes. Table 3 is published in its entirety in the electronic edition; a portion is shown here for guidance regarding its form and content.

^a An integer code is used to describe the filter employed: 1 = V, 2 = I, 3 = R and 4 = B. ^b The image quality index "Q" is described in Section 3.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

The locations of the 23 variable stars are marked in the color–magnitude diagram of Figure 1. We note that many of the variable stars are significantly more red than the comparison stars (0.6 < V − I < 2.3 mag), so that uncertainty in establishing the coefficient c₁ in the equation above can lead to a systematic zero-point error in the resulting magnitude of each variable star.

The light curve of the symbiotic star V1835 Aql is shown in Figure 2. For images obtained within a 12 hr span, the resulting photometry was combined using an error-weighted mean to obtain a single data point per filter per night, shown connected by the green line. The transition times between the different telescope-camera sets are indicated. The photometry appears to be smooth and continuous throughout the time series, indicating that we successfully transformed the different sets of instrumental photometry onto a consistent standard system. Additional tests on several non-variable check stars confirmed that there were no significant jumps at the transitions between data sets that might indicate problems with the photometric calibrations. The ∼100 day gaps in the time series resulted because the star was below an elevation of ∼20° all night between mid-November and late-February each year.

Zoom In Zoom Out Reset image size

Despite the observational scatter, Figure 2 shows fairly regular cyclic behavior with an amplitude of ∼0.9 mag in V and ∼0.3 mag in I. Several points are worth noting. First, the level of maximum light appears to be similar from peak to peak in both filters, though the presence of the data gaps makes this difficult to ascertain. Second, the level of the minima is similar from cycle to cycle, though the star is faint in V at minimum light and the resulting scatter is large. Third, the light curve shape is consistent with that of a contact binary (W Ursae Majoris type variable, EW) or ellipsoidal variable (ELL) with a period of ∼850 days, or with a pulsating variable with a period half that length.

Analysis of the V-band data using the discrete Fourier transform method implemented in VStar Benn (2012) yielded a period 423 of days. A similar application using the I-band data resulted in a period of 431 d, but the lower amplitude in this passband yielded a peak with lower power. We obtained another estimate of the period by estimating the times of minimum light seen in Figure 2 and fitting them with a linear ephemeris of the form ${{JD}}_{\min }=P\,{N}_{\mathrm{cyc}}+{E}_{0}$, to obtain P = 419 ± 10 d and E₀ = 2,456,834 ± 10 days, which we adopt as the best description of the single peak-trough cycle seen in Figure 2.

Figure 3 shows the light and V − I color curves of V1835 Aql phased with twice this period, the expected orbital period if the system is expressing ellipsoidal variation similar to that seen in a sample of ELL-type stars in the Large Magellanic Cloud (LMC) by Soszyński et al. (2004). In our phased light curves, the primary eclipse at ϕ = 0 is marginally deeper (∼1.5σ) than the secondary eclipse. This slight difference is common among ellipsoidal variables with a main-sequence secondary star (Soszyński et al. 2004) and has been attributed to limb-darkening and gravity-darkening in the elongated "tip" of the primary star that points toward the secondary, so that the deeper light curve minimum occurs when this tip is pointing toward the observer; the difference tends to be largest when the primary fills its Roche lobe (Hall 1990).

Zoom In Zoom Out Reset image size

However, the light curve amplitude of V1835 Aql, ∼0.3 mag in I, is large compared with those of ellipsoidal variables in the LCM having periods longer than 500 days (Soszyński et al. 2004), and the amplitude in V is larger still, ∼0.9 mag. Sterken & Jaschek (1996) noted that V-band amplitudes of classical ellipsoidal variables rarely exceed 0.1 mag, with an upper limit of ∼0.4 mag, and their sample light curves show that photometric amplitudes of ellipsoidals are roughly the same size at all visible wavelengths. The differing amplitudes across VRI observed in V1835 Aql can not be explained by simple variations in effective surface area due to viewing a tidally-distorted star at varying orbital phases (Leahy & Leahy 2015). Similarly, the sinusoidally-varying V − I color of V1835 Aql shown in the lower panel of Figure 3 is difficult to explain in a tidally distorted star as described above. We would expect the color to be reddest when the primary star's cool tidal tip is pointed toward us during the deeper light curve minimum, and bluest when it is pointing away during the shallower minimum. Together, the size of the amplitudes and the timing of the color-curve leads us to infer that V1835 Aql is not a simple ellipsoidal variable with an orbital period of 838 days.

The sinusoidal color variations are more in-keeping with a SySt system in which a hot, compact component with a surrounding ionized nebula orbits a cool red giant star. One color cycle would thus correspond to one orbital cycle of P_orb = 419 days, with the object appearing brightest and bluest when the hot nebula is visible; the object would gradually appear fainter and redder as the nebula rotates out of view behind the cool red giant.

Figure 4 shows the star's light and color curves phased with the 419 days period. No sharp-edged eclipses like those seen in Figures 9 and 10 of Gromadski et al. (2013) are evident. This suggests that the orbital inclination is too small for the compact object to pass behind the red giant, or that the luminosity of the compact object relative to the red giant is too low to produce an eclipse (or both). However, the inclination is sufficient for the red giant to partially obscure the luminous nebula during the orbital cycle, thus causing the principal modulation in the light and color curves.

Zoom In Zoom Out Reset image size

Evidence of a very shallow dimming episode is visible at ϕ = 0.5; it appears as a dip in I but only flattens the top of the V and R curves. Modeling the light curves with VStar suggests a sinusoid with P_orb/2 and an amplitude of 0.08 mag matches the light curves in VRI (as shown by the solid curves in Figure 4), leaving the V − I color as an unaffected sinusoid. This behavior is very similar to that seen in the multi-band light curves of YY Her (Mikołajewska et al. 2002) and LT Del (Munari et al. 2021), in which secondary dips of ∼0.1 mag are seen at ϕ = 0.5 in the VRI passbands; their overall range of variation increases in passbands at shorter wavelengths, so that the dip merely flattens the peak of the underlying sinusoid in B and (for YY Her) does not affect it perceptibly in U. In the photometry of another system, T CrB, the secondary minimum is deeper and is clearly visible in B as well as in VRI (Munari et al. 2021), whereas in our photometry of V1835 Aql, the dip is shallower and is clearly visible only in I. In summary, the dips are seen in the redder passbands for all these objects, but the extent to which dips remain visible at shorter wavelengths varies from one object to the next.

Such secondary dips in SySt are widely attributed to ellipsoidal variations caused by the asymmetrical tidal distortion of the red giant star (see Mikołajewska et al. (2002); Munari et al. (2021) and references therein). Because of their primarily geometric cause, the depth of the ellipsoidal dips is roughly constant with wavelength in a given star, as shown for classical ellipsoidal variables in Sterken & Jaschek (1996). Meanwhile, the principle sinusoidal behavior, caused by varying visibility of the hot nebular material over the course of an orbit, tends to be stronger in shorter wavelengths, and thus dominates the ellipsoidal modulation in the blue. The degree to which the irradiated nebula or the red giant photosphere dominates the light curve in a given passband, say V, likely depends on the temperature and luminosity of the nebular material, which in turn depends on the luminosity of the hot, compact stellar component that irradiates it. Munari et al. (2021) and Gromadski et al. (2013) show V-band light curves of a number of SySt that express a range of primary (nebular) and secondary (ellipsoidal) modulation, which indicates a range of luminosities in the compact stars across these systems. Indeed, the systems with stronger nebular emission and larger blue amplitudes have been associated with white dwarfs that fuse the material accreted from the red giant continuously, as soon as it reaches the surface (nuclear burning-type, burn-SySt), whereas the white dwarfs in the systems where the ellipsoidal modulation of the red giant dominates are thought to be powered solely by the gravitational energy of accretion (accreting-only, or accr-SySt), according to Munari et al. (2021) and references therein. Because of the strong nebular modulation seen in V1835 Aql, we place this system among the more commonly-observed burn-SySt subclass. It is intriguing that this system also produced a strong nova-like outburst observed in 1903 (Williams 2005).

Because the secondary, ellipsoidal dip should be more evident at longer wavelengths, we advocate for new time-series observations of V1835 Aql in the near-infrared. Also, obtaining UB photometry over the full orbital cycle would verify our finding that the ellipsoidal modulation is dominated by the nebular emission at theses energies, and enable a more quantitative flux model of the type performed by Mikołajewska et al. (2002) on YY Her. In addition, radial velocities obtained at different orbital phases would help to confirm the 419-d orbital period rather than twice this (ELL variable), and could also lead to mass estimates for the two stars as shown by Mikołajewska (2003a).

The variations of V1835 Aql appear quite regular from cycle to cycle, indicating that variability of the red giant due to chromospheric (starspot) activity or pulsational Mira-like behavior, often seen in symbiotic variables (e.g., Angeloni et al. 2014), is small, ≲0.1 mag. This is consistent with V1835 Aql being an S-type (stellar), rather than the rarer D-type (dusty), symbiotic system (Mikołajewska 2003a). The infrared colors of V1835 Aql listed in Table 2, when corrected for interstellar reddening, are also in good agreement with those of normal M3-4 giants (Kučinskas et al. 2005), indicating that we are viewing the photosphere of the red giant in V1835 Aql rather than its circumstellar dust, and thus it is a SySt of S-type.

We determined the median BVRI magnitudes and light curve extrema from Figure 4, and report them in Table 4, where the brightest and faintest 2% of the data points were eliminated to reduce the effects of outliers and photometric scatter. The column labeled R_f shows the photometric range in filter f as the difference between the faintest and brightest magnitudes listed.

The parallax-based distance for V1835 Aql shown in Table 2 has large uncertainties, indicating that this star remains outside the effective range of Gaia trigonometric parallaxes as of EDR3, though this situation may improve as future data releases become available. We can estimate the distance to V1835 Aql from the V magnitudes in Table 4, the reddening E(B − V) = 0.67 mag (Wehrung et al. 2013), and an absolute magnitude of −0.7 mag based on the M3-4 spectral type (Wehrung et al. 2013). While the uncertainties in these values are large, we estimate that V1835 Aql lies 8–11 kpc from the Sun, 5–6 kpc from the Galactic center and about 0.6 kpc below the Galactic plane. We also note that an upper limit on the distance can be obtained by assuming the red giant fills its Roche lobe; following the arguments in Mikołajewska et al. (2002) for YY Her, one finds that V1835 Aql must lie within ∼15 kpc of the Sun.

In Figure 5, we show the V magnitude as a function of time for several variable stars that display a range of properties. Most of the variables discussed in Wehrung et al. (2013) are labeled "[WLR2013] #" in the AAVSO's Variable Star Index (VSX) using the same identification number listed herein.

Zoom In Zoom Out Reset image size

A table showing the median and extreme light curve values in each bandpass is presented in the Supplement. The extrema were computed at the 2% and 98% points of the magnitude-sorted data sets to reduce the effects of outlier points and photometric scatter. We define the range of variation, R_f , as the difference between the extrema in a particular filter f. In many studies this is termed the amplitude, but since many of the stars studied here have irregular behaviors, we reserve that term for the sinusoids in the Fourier period analysis. As is usually seen for pulsating stars, we found that R_V > R_I for most of the stars in our sample.

Upon inspecting the light curves, it is clear that most of the stars are red, long period variables similar to those observed in the Large Magellanic Cloud by Soszyński et al. (2013). We present in Table 5 the three periods (P₁, P₂, and P₃ in days) and their sinusoidal half-amplitudes (A₁, A₂, and A₃ in mag) with the largest power as determined from the V data for each star using the discrete Fourier transform software VStar (Benn 2012). We excluded from Table 5 periods that are aliases of stronger periods. The semi-amplitudes are closely correlated with the power metric output by VStar, so larger amplitudes indicate higher power and more likely periods. We include the range, R_V , in the second column of Table 5 for comparison with the amplitudes. In some cases, a light curve modeled using the three highest-power sinusoids does not account for all the variation seen in the observed light curve, indicating additional sources of natural variation that may include more periods operating in the star, the presence of truly irregular (chaotic) pulsation, magnetic starspot activity, convective cells, or absorption by varying amounts of circumstellar dust. Other stars, like #27, have a single period that dominates the power spectrum. The periods listed in bold are the ones we find are most successful in characterizing each star's light curve.

In the following subsections, we highlight several stars in each of four broad categories. A detailed discussion of each star, along with V-band light curves of stars not shown in Figure 5, are presented in the Supplement.

5.1. High Amplitude, Regular LPVs

5.2. Lower Amplitude, Multi-periodic LPVs

5.3. Bluer, Non-LPV Variables

5.4. Variables in the Wider Fields

We obtained time-series CCD photometry in the BVRI passbands of the recently identified symbiotic nova V1835 Aql over an interval of 5.1 yr. The star varies through a magnitude range of R_V ≈ 0.9 mag and R_I ≈0.3 mag with a regular cycle described by the ephemeris ${{JD}}_{\min }=(419\pm 10)\,{N}_{\mathrm{cyc}}+(2,456,834\pm 10){\rm{d}}$. This orbital period is relatively short for a SySt, ranking eighth of 53 objects in the distribution given by Gromadski et al. (2013). No sharp-edged stellar eclipses are evident. We see a dip at ϕ = 0.5 in the light curve shown in Figure 4 that we attribute to tidal distortion of the giant star and the consequent ellipsoidal effect on the light curve. This hints that V1835 Aql may transfer some of its mass to the hot component via Roche lobe overflow (Mikołajewska 2003b), rather than the more widely-accepted model of stellar wind transfer. We show that any intrinsic variation of the red giant, for example caused by pulsation, is weak; this, and the dereddened near-infrared color of the red giant, are consistent with V1835 Aql being a S-type (stellar) SySt. The strong photometric modulation of the object in the bluer passbands suggests that the compact stellar object that irradiates the giant star's wind has a relatively high luminosity, consistent with fusion of accreted hydrogen continually as it reaches the stellar surface: a burn-SySt. It is noteworthy that V1835 Aql is among a small number of SySt that have been observed to undergo a major nove-like outburst (Williams 2005).

We also provide BVRI time series photometry on 22 variable stars in the field around V1835 Aql, nine of which were initially detected in Wehrung et al. (2013). We list the stars' photometric properties and periods found using Fourier analysis. Most turn out to be long period variable red giants, including one Mira, four semi-regular (SRA), nine multi-periodic (SRB), and five LPVs which showed convincing periodic behavior for which we did not have enough data to classify with confidence (SR). Many of these stars have long secondary periods (LSPs) and two stars appear to have two LSPs each, which seems viable under the hypothesis that the LSP behavior is caused by dimming due to dust clouds entrained by orbiting, low-mass companion objects (Soszyński et al. 2021). Detailed analysis of each star and light curves are provided in the Supplement.

We recommend that V1835 Aql receive additional photometric observations in the near-ultraviolet and near-infrared regions (e.g., UB, JHK) in order to confirm the ellipsoidal variations suspected herein, and to quantify its physical properties. Spectroscopic radial velocity measurements over all phases of the light curve would also confirm the orbital period and perhaps provide a mass ratio as shown in Mikołajewska (2003a) We also note that many of these stars, including V1835 Aql, are receiving ongoing optical photometric monitoring by the Zwicky Transient Facility (Bellm et al. 2019) and the Gaia astrometric satellite (Gaia Collaboration 2016).

This work represents partial fulfillment of a Master of Science degree in Physics at BGSU for RVC. We thank BGSU undergraduate Bianca Legeza for help in obtaining and analyzing data from online sources. We thank Howard Bond and an anonymous reviewer for helpful feedback on our manuscript. This research was made possible through the use of the AAVSO Photometric All-Sky Survey (APASS), funded by the Robert Martin Ayers Sciences Fund, as well as the AAVSO's Variable Star Index (VSX). We thank the Robert Martin Ayers Science Fund for sponsoring our observing time on the PROMPT telescopes for this project. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This research utilized the Image Reduction and Analysis Facility (IRAF), which was developed and distributed by NOAO through AURA, Inc. under agreement with the NSF.

Facilities: AAVSO - , CTIO:PROMPT - .

Software: IRAF (https://iraf.net/), VStar (https://www.aavso.org/vstar).