Peremennye Zvezdy (Variable Stars) 45, No. 15, 2025 Received 28 October; accepted 6 November.	Article in PDF
	DOI: 10.24412/2221-0474-2025-45-154-162

Near-infrared Photometry and Dust Envelope Modeling of the Hot Variable Post-AGB Star IRAS 21546+4721

N. P. Ikonnikova, M. A. Burlak, B. S. Safonov

Sternberg Astronomical Institute, Lomonosov Moscow State University, Universitetskij pr. 13, Moscow 119234, Russia

1. Introduction

Post-AGB objects are stars in the final stage of their evolution after the asymptotic giant branch (AGB) phase. They represent an intermediate stage between the AGB phase and the white dwarf stage. After the superwind phase ceases at the tip of the AGB, the star possesses a CO core with a mass of 0.5-0.8 , surrounded with an extensive gaseous envelope that mimics characteristics (luminosity, surface gravity) of a supergiant. The object is also enveloped with a dust shell, which becomes more transparent as it expands. The star gradually contracts, its temperature rises, and at  K, ionization of the surrounding envelope begins, leading to the appearance of a low-excitation emission spectrum from the nebula.

IRAS 21546+4721 ( , , 2000) is a hot post-AGB star in the early stage of planetary nebula formation. Its spectrum exhibits both stellar absorption lines and emission lines from the surrounding gaseous envelope (Suárez et al. 2006; Ikonnikova et al. 2025). The presence of a dust envelope is evidenced by infrared emission (Manchado et al. 1989; Venkata Raman et al. 2017). Recently, we detected rapid, night-to-night, non-periodic optical variations with an amplitude of approximately 03 in all bands, as well as either a reddening of the index during phases of increased brightness or the absence of a statistically significant correlation between color indices and brightness (Ikonnikova et al. 2025).

This paper presents near-infrared (NIR) photometric observations of the star in the bands. We construct its spectral energy distribution and derive physical parameters of the dust envelope.

2. Observations and reductions

Our photometry was obtained in 2021-2022 with the ASTROnomical Near-InfraRed CAMera (ASTRONIRCAM) (Nadjip et al. 2017) mounted on the 2.5-m telescope of the Caucasian Mountain Observatory of the Sternberg Astronomical Institute, Lomonosov Moscow University. The observations were carried out in dithering mode to acquire images in the bands of the Mauna Kea Observatories Near-Infrared (MKO-NIR) system (Simons and Tokunaga 2002). Each frame was automatically reduced and calibrated using the pipeline described in detail by Tatarnikov et al. (2023). The standard reduction procedures included correction for non-linearity, bad pixels, and dark subtraction. Flat-fielding and background subtraction were performed using custom Python scripts. Instrumental magnitudes for the target star were derived using differential aperture photometry relative to a primary reference star. Two additional nearby stars were used to monitor the photometric stability of the reference star (Table 1). The 2MASS magnitudes of these comparison stars were transformed to the MKO-NIR system using the equations from Leggett et al. (2006). The -band magnitudes were estimated from the 2MASS and magnitudes using the transformation equations derived in Tatarnikov et al. (2023). For each filter and pointing, multiple frames were obtained; the reported magnitudes are based on the mean of these measurements, and the uncertainties were calculated as the standard deviation of individual values. The average photometric uncertainties are: , , , . Ten individual -band brightness measurements of the star are listed in Table 2. Photometry in the bands is provided in both the MKO-NIR and 2MASS systems in Table 3.

3. Data analysis

Figure 1 shows the light curves of IRAS 21546+4721 in the  bands, along with the -band light curve from Ikonnikova et al. (2025). Although the vertical scale in the plot makes the variations appear small, the star actually varies in brightness with amplitudes up to 03 in all bands.

Although no systematic NIR monitoring of hot post-AGB stars has been conducted to date - with only individual brightness estimates available, - optical variability has been detected for 17 such objects, as listed in Ikonnikova et al. (2025). It is characterized by irregular fluctuations with maximum amplitudes of 02-04, typical timescales ranging from a fraction of a day to several days, and no detectable periodicity.

There is still no consensus on the physical mechanisms driving the variability of hot post-AGB stars. The primary hypotheses under consideration attribute it to an unsteady stellar wind and pulsations (Handler 1999). It is possible that both processes contribute: the gradual brightness changes may be caused by variations in the mass loss rate, while the short-term fluctuations are most likely due to pulsations. The detection of pulsation periods probably requires better time-resolution observations.

Table 4 presents NIR () photometric data from the literature together with our averaged values. The data from García-Lario et al. (1997) and the 2MASS catalogue include formal measurement uncertainties. In contrast, our values represent mean magnitudes over the entire observational series, uncertainties given as the root-mean-square (RMS) deviations, which reflect the object's intrinsic variability. The RMS uncertainties of our measurements are comparable to the photometric errors quoted in the 2MASS catalog.

The study by García-Lario et al. (1997) provides brightness estimates for two observational epochs, which show slight discrepancies. While this may hint at intrinsic variability, the data are affected by substantial measurement uncertainties. Notably, the García-Lario et al. (1997) values differ substantially from the 2MASS data - even when accounting for their large error bars, - most likely due to differences between the photometric systems of the instruments used. In contrast, a comparison of our data with the 2MASS results (which are tied to the same MKO-NIR photometric system after transformation) reveals discrepancies that can be attributed to the object's genuine variability.

Figure 2 shows a () vs. () color-color diagram, where a clear difference is seen between the 2MASS data and our modern measurements. An anomaly is observed: the () color index has become bluer, while the () index has turned redder. If the long-term variability were due to dust-shell evolution, both indices would have shifted coherently - either redward with increasing optical depth or blueward with decreasing optical depth. A similar coherent trend would be expected for changes in stellar temperature: decreasing temperature would cause both color indices ( and ) to shift redward, increasing temperature would lead to a blueward shift. For post-AGB stars, theory predicts a monotonic rise in temperature over time. The observed opposing trends in the color indices currently lack a definitive explanation.

4. The model of the circumstellar dust envelope

The star exhibits a significant infrared excess due to the presence of a circumstellar dust shell. Our goal was to model the star's SED across a broad wavelength range (from 0.3 to 60 m) and to determine parameters of the dust shell.

We employ numerical Monte Carlo Radiative Transfer (MCRT) calculations using the RADMC-3D software (Dullemond et al. 2012) as our model. This code simulates the emission, propagation, scattering, and absorption of individual photon packets. It accounts for wavelength-dependent absorption coefficients and angle-dependent scattering matrices. Thermal equilibrium is maintained following the approach of Bjorkman & Wood (2001). Escaping photons are recorded, and their properties are used to construct the SED. Interpretation of observational data for this study was performed using webMCRT¹ - a web-based interface to RADMC3D. Relevant model links are provided throughout the text for reference.

Our modeling is based on our data, which represent averaged values obtained from our observations with the RC600 telescope between 2020 and 2024, as published in Ikonnikova et al. (2025) and our data (Table 5) as well as the data available from different catalogues listed in Table 6. For absolute calibrations, we adopted the following sources: for bands, Straizys (1992); for bands, Bessell (1979); for bands, Koornneef (1983).

The observed SED has a shape that cannot be reproduced using a single envelope described by a simple density distribution law of the form . Therefore, we defined the envelope as three nested layers, fixing the power-law index in each layer. This assumes that the envelope was formed through several episodes of mass loss, each occurring at a constant mass loss rate.

In accordance with the principle of parsimony, our subsequent modeling employed a three-shell composite dust envelope. This model was built upon the following foundational assumptions:

• the distance to the star was adopted as 9781 pc (Bailer-Jones et al. 2021), based on the parallax from GAIA DR3 (Gaia Collaboration, 2021);

• the stellar radiation was modeled using a synthetic spectrum for a B0 supergiant from Pickles (1998);

• the interstellar extinction was adopted to be .

IRAS 21546+4721 is classified as a carbon-rich object. This is determined by the presence of fullerenes (C) and other carbon compounds, such as polycyclic aromatic hydrocarbons (PAHs), which have been detected in its spectra. In particular, the identification of vibrational modes of C fullerene indicates carbonaceous features in the star's circumstellar environment (Venkata Raman et al. 2017). To minimize ambiguity in the absence of specific compositional data, the dust was modeled as amorphous carbon, using the optical constants from Suh (2000).

Table 7 lists the parameters of the shell components, including the inner and outer radii, dust grain radius, optical depth, and dust masses. Figure 3 represents the best-fitting model SED with minimal normalized deviations from the observed data. The components of the dust shell are numbered in Table 7 and Fig. 3 in order of increasing distance from the star. The most massive and most distant shell (3) was formed during the AGB phase as a result of mass loss from thermal pulses. The other two shells, (1) and (2), formed later in the post-AGB phase. It is likely that the innermost shell (1) is still being replenished with material. This is because, as the star's temperature increases during the post-AGB stage, its stellar wind intensifies.

The observed SED is reasonably well described by our model, with the exception of the 5-10 micron wavelength range. Here, the model fails to reproduce the strong PAH bands seen in the observed spectrum, as it is limited to simulating the continuum.

The study by Venkata Raman et al. (2017) presented Spitzer mid-IR spectra for a large sample of post-AGB stars, including IRAS 21546+4721. Based on the presence of PAH features at 6.2, 7.7, and 11.2 m in its spectrum, this star was classified as a PAH-type object. Furthermore, its SED was modeled using the DUSTY software (Ivezic et al. 1999). The dust grains in the model were composed of graphite (Gr). The stellar temperature was adopted as 10,000 K, although more recent data indicate that the star is significantly hotter, at approximately 24,000 K (Ikonnikova et al. 2025). The resulting model consists of two spherically symmetric shells. A density distribution falling off as was used for both shells. It is assumed that PAHs and dust exist in two separate shells, with the PAHs located closer to the star than the dust grains.

5. Conclusions

We present the results of our NIR photometric monitoring of the hot post-AGB star IRAS 21546+4721, conducted in the bands during 2021-2022 with the 2.5-m telescope of the Caucasian Mountain Observatory. The main results of our study are as follows:

(1) The star exhibits non-periodic brightness variations with amplitudes up to 03 in all NIR bands, consistent with the behavior previously observed in the optical domain. This confirms that IRAS 21546+4721 is a rapidly variable hot post-AGB star.

(2) A comparison of our NIR data to earlier 2MASS measurements reveals significant long-term changes in the color indices: the color has become bluer, while the color has reddened. This opposing behavior cannot be explained by simple changes in the dust optical depth or stellar temperature alone and remains unexplained.

(3) The SED of the star from 0.35 to 60 m was successfully modeled using a three-component dusty envelope with a power-law density distribution ( ). The model assumes amorphous carbon dust, consistent with the carbon-rich nature of the object, and reproduces the observed IR excess.

(4) The best-fitting model consists of three nested shells with inner radii of 5, 10, and 1500 au, and outer radii of 100, 500, and 15000 au, respectively. This multi-layered structure suggests multiple episodes of mass loss during the star's evolution.

In summary, IRAS 21546+4721 represents an important laboratory for studying the late stages of stellar evolution, the dynamics of circumstellar envelopes, and the onset of planetary nebula formation.

Acknowledgments: We are especially grateful to the CMO staff for maintaining the equipment and helping in observations. The work of N.P. Ikonnikova (data analysis and article writing) and B.S. Safonov (dust shell modeling) was supported by the RSF grant 23-12-00092.

References:

Bailer-Jones. C. A. L., Rybizki, J., Fouesneau, M., et al. 2021, Astron. J., 161, No. 3, article id. 147

Bessell, M. S. 1979, Publ. Astron. Soc. Pacific, 91, 589

Bjorkman, J. E. & Wood, K. 2001, Astrophys. J., 554, No. 1, 615

Cutri, R. M., Wright, E. L., Conrow, T., et al. 2013, VizieR Online Data Catalog, II/328

Dullemond, C. P., Juhasz, A., Pohl, A., et al. 2012, Astrophysics Source Code Library, ascl:1202.015

Gaia Collaboration, Brown, A. G. A., Vallenari, A., Prusti, T., et al. 2021, Astron. & Astrophys., 649, id. A1

García-Lario, P., Parthasarathy, M., De Martino, D., et al. 1997, Astron. & Astrophys., 326, No. 3, 1103

Handler, G. 1999, Astron. & Astrophys., Suppl. Ser., 135, No. 3, 493

Helou, G. & Walker, D.W. 1988, Infrared astronomical satellite (IRAS) catalogs and atlases, Vol. 7, pp. 1-265

Houck, J.R., Roellig, T.L., van Cleve, J., et al. 2004, Astrophys. J., Suppl. Ser., 154, No. 1, 18

Ikonnikova, N., Burlak, M., & Dodin, A., 2025, Galaxies, 13, No. 2, 31

Ivezic, Z., Nenkova, M., & Elitzur, M. 1999, User Manual for dusty, University of Kentucky Internal Report, accessible at http://www.pa.uky.edu/moshe/dusty

Koornneef, J. 1983, Astron. & Astrophys., 128, No. 1, 84

Leggett, S. K., Currie, M. J., Varricatt, W. P., et al. 2006, Monthly Notices Roy. Astron. Soc., 373, No. 2, 781

Manchado, A., Pottasch, S. R., García-Lario, P., et al. 1989, Astron. & Astrophys., 214, No. 1/2, 139

Murakami, H., Baba, H., Barthel, P., et al. 2007, Publ. of the Astron. Soc. of Japan, 59, issue sp2, S369

Nadjip, A. E., Tatarnikov, A. M., Toomey, D. W., et al. 2017, Astrophys. Bull., 72, No. 3, 349

Pickles, A. J. 1998, Publ. Astron. Soc. Pacific, 110, 863

Simons, D. A. & Tokunaga, A. 2002, Publ. Astron. Soc. Pacific, 114, 169

Straizys, V., 1992, Multicolor stellar photometry. Tucson: Pachart Publishing House

Suárez, O., García-Lario, P., Manchado, A., et al. 2006, Astron. & Astrophys., 458, No. 1, 173

Suh, K.-W. 2000, Monthly Notices Roy. Astron. Soc., 315, No. 4, 740

Tatarnikov, A. M., Zheltoukhov, S. .G., Shatsky, N. I., et al. 2023, Astrophys. Bull., 78, No. 3, 384

Venkata Raman, V., Anandarao, B. G., Janardhan, P., & Pandey, R. 2017, Monthly Notices Roy. Astron. Soc., 470, No. 2, 1593

---