Highlight
Free Access
Issue
A&A
Volume 553, May 2013
Article Number L1
Number of page(s) 4
Section Letters
DOI https://doi.org/10.1051/0004-6361/201220715
Published online 26 April 2013

© ESO, 2013

1. Introduction

In the course of the formation of a star, an equatorial disk is present, whose purpose evolves from an angular momentum re-distributor, which facilitates star growth, to a planet builder. Dust grains in the disk change chemically (Bouwman et al. 2001) and physically through growth (van Boekel et al. 2004) and collisions (Wyatt et al. 2007b). Observations demonstrate on the one hand the presence of dust disks by their tell-tale thermal infrared (IR) spectrum, whereas on the other hand hundreds of mature planetary systems are now known. But how the dust disk evolves to a planetary system is poorly understood as yet (Williams & Cieza 2011).

RZ Psc is a solar-type star (K0 IV, Herbig 1960) and well known for its brightness variability with time. The variability has all the hallmarks of the so-called UXOR variability seen among pre-main sequence stars. They sporadically have photometric minima with amplitudes of ΔV ≈ 2m−3m that last from a few days up to a few weeks. During a minimum, the UXOR displays bluer optical colours and an increased linear polarization due to an increased contribution by scattered light off small dust grains (e.g. Grinin 1988; Grinin et al. 1991). This type of variability is strictly associated with the occultation of the star by dust in the optically thick accretion disks of stars younger than 10 million years (Dullemond et al. 2003). RZ Psc is estimated to be approximately three times as old (Grinin et al. 2010a,b, hereafter Paper I). Correspondingly, the star does not display the benchmark properties of young stars such as ionized gas transitions or excess emission by hot (1500 K) dust (Bertout 1989; Paper I). Rz Psc is therefore enigmatic because of a variability that is normally caused by optically thick accretion disks. In this follow-up study, we report that the object has one of the strongest infrared excesses observed to date, and we provide a detailed periodicity analysis of the optical variability.

2. Observational data

To shed light on the cause for the star’s continuum emission variability, which is peculiar for its age, we explored the mid-infrared and far-infrared wavelength region by means of data obtained by a variety of infrared satellite missions and ground-based surveys, viz. the Wide-field Infrared Survey Explorer (WISE, Wright et al. 2010), the Infrared Astronomical Satellite (IRAS), AKARI (Ishihara et al. 2010), and the Two Micron All Sky Survey (2MASS, Cutri et al. 2003). We also present new V-band photometry taken at the Crimean Astrophysical Observatory (CrAO) that extend the star’s light curve to March 2012 (see Fig. 2).

Representative UBVRI magnitudes for the star’s photosphere are taken from the mean value of the three brightest measurements, assuming that these are least affected by circumstellar extinction. The 2MASS near-IR magnitudes in the JHK bands reveal no change with respect to the measurements by Glass & Penston (1974), H = 9.7 ± 0.2 and K = 9.78 ± 0.1. The WISE point source coincides with RZ Psc within 0.3′′ and is the only catalogued WISE source within a 23′′ radius of our target. We also note that the IRAS 12 μm flux is consistent with the WISE flux measurement at 11.6 μm. Finally, the IRAS Faint Source Reject Catalog contains flux upper limits at 60 μm and 100 μm.

3. Results and discussion

In Paper I, we estimated an age for RZ Psc of 30−40 Myr based on its kinematics and the lithium absorption. An age constraint using the star’s space motion was performed, exploiting the high galactic latitude of b ~ 35°, and assuming that the star formed at b = 0°. Clearly, this is an uncertain exercise, but the result is consistent with the observed Li equivalent width, which indicates an age between 10 and 70 Myr. Recently, we measured a vrad of − 2.0 ± 1.5   km   s-1 (Potravnov & Grinin, in prep.), consistent with the velocities reported in Shevchenko et al. (1993), but in disagreement with the literature value (− 11.5   km   s-1) used in Paper I. The new value changes the kinematic age slightly to tkin ~ 25 ± 5 Myr (standard error). Arguments in support of a system older than 10 Myr include the absence of interstellar dust extinction towards RZ Psc (the star lies far from active star formation regions) and the lack of Hα emission. In the following, we discuss the system’s properties assuming that the star has finished its formation process, because the available age constraints exceed the characteristic time for optically thick accretion disks found for classical T Tauri stars, but we keep in mind that the system is nonetheless relatively young.

thumbnail Fig. 1

SED of RZ Psc. A K0 IV model photosphere (Pickles 1998) fits the optical and near-IR measurements shortward of 3 μm. Longward of 3 μm, the excess emission is markedly dominant and contributes 8% to the total source luminosity. The solid black line corresponds to the total flux of the stellar photosphere plus a single black-body curve of 500 K. Some uncertainties in the flux measurements are smaller than the plot-symbol. The errorbars in wavelength indicate the width of the filter, which corresponds to 50% of the transmission.

Open with DEXTER

3.1. Spectral energy distribution

The resulting spectral energy distribution (SED) is presented in Fig. 1. It reveals the star’s photospheric emission at visual wavelengths and strong excess emission that dominates the total radiation from 3 μm onwards. The photospheric fit does not require an extinction correction (Kaminskiĭ et al. 2000). IR excess emission is observed among a small fraction of main-sequence stars, viz. the debris disk objects (Wyatt 2008). The emission is caused by warm dust particles released in collisions between planetesimals in belts similar to the asteroid and Kuiper belt of our solar system (Wyatt 2008). However, the excess emission observed in RZ Psc is found to have two outstanding properties that set it apart from regular debris disk systems.

  • 1.

    The dust emission is well approximated by asingle 500 K Planck function,strongly constrained by theWISE 22 μm measurement. The temperature of the warm dust is far lower than the dust sublimation temperature of approximately 1500 K. This implies that (1) orbits closer to the star are (practically) devoid of dust grains, otherwise dust excess emission would be detected in the 2 μm wavelength region; and (2) that the dust distribution is limited in radial extent (a ring rather). Assuming heating by stellar irradiation and dust particles in thermal equilibrium with an optically thick environment, one can estimate the radial distance of the dust by applying Td ≈ Teff(r/r)− 1/2. Adopting the stellar parameters Teff = 5250 K and R = 1.5  R (Paper I) delivers a characteristic distance for the dust of 0.7 AU. Dropping the optically thick assumption and assuming optically thin dust instead, this characteristic distance is reduced to 0.4 AU. In addition, a cold (100 K) dust component cannot be excluded by the current set of measurements, but the total emission at 60 μm should be ≲0.05 Jy.

  • 2.

    The second unusual property is the high fractional contribution of the excess to the total luminosity, which amounts to 8%; this excess places the star among the non-accreting objects with the strongest IR excesses1. We can estimate the expected fractional contribution by the production of small dust particles in a model for a steady-state, collisional cascade of an asteroid belt, which is thought to be valid for debris disks (Wyatt et al. 2007a). In brief, a steady state model adopts a planetesimal size distribution that does not evolve, except that the largest-sized bodies disappear by collisions and the minimum size is set by stellar radiation pressure effects. The cascade model predicts a maximum IR belt luminosity with time (Eq. (18) in Wyatt 2008). If the evolution of the planetesimal belt of RZ Psc is similar to that observed in A-type star debris disks (for which a fitting steady-state representation is parametrized by a maximum asteroid size of 60 km, planetesimal strength of 150 J/kg, and eccentricity of 0.05), then with a characteristic distance of 0.7 AU a belt width of 0.3 AU, a fractional luminosity contribution of a few times 10-5% is predicted at the adopted age of RZ Psc. This is a common value for debris disks. Much more mass in small dust grains needs to be produced in RZ Psc than can be accounted for by the steady-state collisional paradigm. We can use Eq. (4) from Wyatt (2008) to convert the 8% fractional contribution to a lower limit on the mass, viz. Mdisk > 1 × 1023 g. This conversion assumes that the fractional luminosity defines the effective cross-sectional area of the dust, i.e. an optically thin environment (dust edge at 0.4 AU). Additionally, we adopt a single size and density for the dust particles of 10  μm and 3.3  g   cm-3 (following Lawler & Gladman 2012). However, if the same dust is responsible for the excess emission and the UXOR variability, then clearly the size of the dust should be (sub-)micron sized.

A dust belt with a sharp inner edge can be produced by dynamical interactions between the dust and a secondary object located within the gap between star and belt (Williams & Cieza 2011; Kraus & Ireland 2012). If we assume this to be the case in RZ Psc, then we can derive the following about the secondary’s properties. For a zero eccentricity orbit, the semi-major axis of the secondary has to be approximately half that of the inner edge of the belt, (semi-major axis between 0.2 and 0.4 AU) (Artymowicz & Lubow 1994). The absence of any radial velocity variation down to 2  km   s-1 (Shevchenko et al. 1993) results in a secondary mass of ≲38  MJup for optically thin dust (or ≲53  MJup for the optically thick case). For a non-zero eccentricity orbit of a secondary object sculpting the inner rim of the dust distribution, the semi-major axis could be even smaller (e.g. Beust 2003) and the mass upper limits are then also correspondingly lower.

Alternative processes for explaining an inner gap by means of grain growth (as the first step in the planet formation process) or photo-evaporation (responsible for the dissipation of the primordial accretion disk, especially the disk’s outer-parts) are unlikely because they play a role in optically thick primordial disks during the first 10 Myr of the evolution (Cieza 2008).

thumbnail Fig. 2

Light curve of RZ Psc over 40 years. More than one thousand measurements are plotted (small blue stars). Gaps in the measurement coverage are caused by the seasonal observability of the object. The deepest minima (ΔV > 1) are marked with open circles. Additionally, the flux at bright phases varies periodically. This periodicity is revealed by sigma-clipping and time-binning the data, resulting in the points represented by black circles, whose 12.44-year period is marked by the sinusoid (see text). Typical uncertainties in the clipped-binned data are indicated only in 10% of the plotted points.

Open with DEXTER

3.2. Optical variability

A large mass in small dust particles provides a reservoir for the dust occultation events depicted in Fig. 2. The figure shows the photometric activity of the star during the past forty years. The light curve is based on our own observations and data from the literature (Zajtseva 1985; Kardopolov et al. 1980; Pugach 1981; Kiselev et al. 1991; Shakhovskoi et al. 2003). Two phenomena can be identified:

  • (1)

    Brightness decreases of up to 2.5 visual magnitudes.They occur on average once every year, but the events areaperiodic. The brightness minima last 1to 2 days. From the rateof flux change, one can estimate the tangential veloc-ity and approximate the distance for an opaque screen(Paper I). From this, an orbital dis-tance of 0.6 AU is found,not inconsistent with the distance estimated fromthe excess emission. The flux minima are accompa-nied by an increased degree of polarization (Kiselevet al. 1991; Shakhovskoiet al. 2003), which, taken together,unambiguously identify small, micron-sized dust grains as thecause (Grinin et al. 1991). Theoccultation time then limits the characteristic size for the clumpsto ~0.05 AU. The extinction at the beginning of the eclipses displayed by RZ Psc is typical for UX Ori stars, and these can be described approximately by assuming MRN size distribution for sub-micron sized particles (Voshchinnikov et al. 1995). Adopting the opacity at optical wavelengths for an ISM dust mixture (Natta & Whitney 2000), and with the observed τv = 3 of the dust clump (a common value for UXORs), we derive a mass of a few times 1020 g (assuming the clump consists of dust only).

  • (2)

    A modulation of the peak flux with a cycle of 12.4 years. The brightness variability of RZ Psc has been presented in previous publications that reported a quasi-periodic variation of the stellar flux level during bright periods, albeit with different periods (Shakhovskoi et al. 2003; Rostopchina et al. 1999). Here we present a rigorous analytical period search with the aim to settle the question of any periodicity in RZ Psc’s photometric data of the past forty years.

The light curve data present two challenges for period searches. First, the data sampling is unevenly spaced in time both because of seasonal observability and weather conditions. Second, the brightness variability is dominated by an intrinsically irregular component that dominates the variability on a time-scale of days, i.e. the obscuration of the star by the dust clumps. The amplitude of the irregular variation is larger than the measurement precision of less than approximately 0.03 magnitudes. We performed a period search on a subset of the data, which was defined by discarding measurements deviating more than 2σ from the median brightness of the source. This step aims at removing the signal caused by the obscuration events that lead to the deep brightness minima. In the next step, the data were averaged per time-bin with the aim to average out the small amplitude of ~0.3m daily variability of the source. An optimal bin size of 29 days was objectively calculated according to a method that minimizes the bin size by taking into account the bin statistics (mean and variance) of the finite sample (Shimazaki & Shinomoto 2007). The bin-averaged data points are represented in Fig. 2 by blue symbols and the errorbars represent the uncertainties taken to be the standard deviation per bin. If a time bin contains only one measurement, the average time bin standard deviation is assigned. Finally, a period search was performed by computing the generalized Lomb-Scargle (GLS) periodogram (Zechmeister & Kürster 2009). This methodology revealed a significant power at a period of 12.34 years with a false-alarm probability (FAP) of 6.8 × 10-4. No other significant power peaks are present in the GLS periodogram. A subsequent sinusoidal fit delivers a refined period of 12.44 years and an amplitude of 0.5m, which is presented in Fig. 2. Removing of this periodic signal from the data results in a power-spectrum without significant power beyond the noise.

The cyclic variability of cool stars can be connected with the magnetic surface activity (magnetic cycles, see, e.g. Grankin et al. 2008). This type of activity is frequently accompanied with the rotational brightness modulation of stars that are not viewed pole-on (e.g. Vrba et al. 1988). RZ Psc is observed close to equator-on and its brightness does not display any rotation modulation. The cause for the long-term cyclic variability of RZ Psc could be a warped disk (Grinin et al. 2010a,b) or large-scale perturbations in a disk, as is suggested for UXORs (Grinin et al. 1998). Such perturbations can be caused by the orbital motion of a co-planar low-mass companion (Demidova et al. 2010). We speculate that if this is the case, then the system is required to have a second low-mass companion. Taking into account a 30% uncertainty in star mass and 0.25 year on the period, the semi-major axis of this tertiary component is a = 5.3 ± 0.6 AU. Applying this particular model is possible only if the outer part of the disk (beyond the orbit of the second companion) contains some amount of gas and fine dust. This material could be the remnant of the primordial disk. This scenario would predict a far IR excess with T ≤ 100 K. We underline that for a full understanding of the RZ Psc system, in particular the effects on the dust distribution by the two inferred companion objects, numerical modelling is required.

thumbnail Fig. 3

Generalized Lomb-Scargle periodogram and window function. The computed periodogram (upper plot) is a convolution of the astrophysical signal and the sampling function (window function, lower plot). The periodogram is obtained from the sigma-clipped, bin-averaged dataset. The horizontal dashed lines indicate FAP levels of 10–1, 10–2, and 10–3. There is one significant peak at a period of 12.34 years with an FAP of 6.8 × 10-4. The window function shows a significant peak at 365 days that is caused by the seasonal visibility of the source. Significant alias peaks caused by the sampling period are absent in the GLS periodogram.

Open with DEXTER

3.3. Warm debris disks

RZ Psc adds to a growing number of main-sequence stars with exceptionally strong, warm (500 K) IR excess. The excesses are orders of magnitude stronger than can be explained by the collisional cascade model that is applicable to regular debris disks (Song et al. 2005; Melis et al. 2010; Zuckerman et al. 2012; Lawler & Gladman 2012). Evidence is mounting that the warm debris disks are generated by transient or stochastic events based on the observed altered emission properties of the dust grains (Olofsson et al. 2012). Proposed transient events constitute a recent collision between two major bodies (such as rocky planetary embryos or even planets, Melis et al. 2010) or a second planetesimal belt at larger radii that feeds the inner dust distribution with evaporating comets and/or induces collisions, partially analogous to the solar system Kuiper belt (Olofsson et al. 2012). Involved time scales would favour the former explanation, which is moreover supported by the fact that the warm dust phenomenon near solar-type stars may occur only in the first 30 to 100 Myr, as judged from a sample of four field stars (Melis et al. 2010). Furthermore, young stellar clusters indicate a maximum in the number of stars that show 24 μm excess emission around approximately 40 Myr (Smith et al. 2011). RZ Psc is unique in the sense that the system’s near equator-on orientation and dust occultations give important clues on the dynamics of the system.

4. Conclusions

We have detected a strong mid-IR excess in the young star RZ Psc with an 8% fractional contribution to the bolometric luminosity of the system. The excess traces warm dust and is well described by a single Planck function of ~500 K. This suggests a radially compact dust distribution (a ring), with the inner-edge at 0.7 AU if the dust is optically thick. We also found a periodicity of 12.4 years in the maximum optical brightness of the object. Although uncertain, the estimates for the age indicate that the star is beyond a formative phase and the mid-IR excess is unlikely to be caused by a primordial disk. Therefore, copious amounts of small dust need to be continually produced in this system to explain the optical occultation events (UXOR phenomenon), providing a dynamical view on the dust production process. The RZ Psc system could play a key role in understanding the transition from primordial to debris disks.


1

A contribution >16% was recently reported for the 60 Myr solar-type star V488 Per (Zuckerman et al. 2012).

Acknowledgments

This publication makes use of data products from the Wide-field Infrared Survey Explorer, which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory/California Institute of Technology, funded by the National Aeronautics and Space Administration, and it makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation. This research is based on observations with AKARI, a JAXA project with the participation of ESA and has made use of NASA’s Astrophysics Data System Bibliographic Services. V.G. and I.P. were supported by grant of the Presidium of RAS P21 and grant N.Sh.-1625.2012.2.

References

All Figures

thumbnail Fig. 1

SED of RZ Psc. A K0 IV model photosphere (Pickles 1998) fits the optical and near-IR measurements shortward of 3 μm. Longward of 3 μm, the excess emission is markedly dominant and contributes 8% to the total source luminosity. The solid black line corresponds to the total flux of the stellar photosphere plus a single black-body curve of 500 K. Some uncertainties in the flux measurements are smaller than the plot-symbol. The errorbars in wavelength indicate the width of the filter, which corresponds to 50% of the transmission.

Open with DEXTER
In the text
thumbnail Fig. 2

Light curve of RZ Psc over 40 years. More than one thousand measurements are plotted (small blue stars). Gaps in the measurement coverage are caused by the seasonal observability of the object. The deepest minima (ΔV > 1) are marked with open circles. Additionally, the flux at bright phases varies periodically. This periodicity is revealed by sigma-clipping and time-binning the data, resulting in the points represented by black circles, whose 12.44-year period is marked by the sinusoid (see text). Typical uncertainties in the clipped-binned data are indicated only in 10% of the plotted points.

Open with DEXTER
In the text
thumbnail Fig. 3

Generalized Lomb-Scargle periodogram and window function. The computed periodogram (upper plot) is a convolution of the astrophysical signal and the sampling function (window function, lower plot). The periodogram is obtained from the sigma-clipped, bin-averaged dataset. The horizontal dashed lines indicate FAP levels of 10–1, 10–2, and 10–3. There is one significant peak at a period of 12.34 years with an FAP of 6.8 × 10-4. The window function shows a significant peak at 365 days that is caused by the seasonal visibility of the source. Significant alias peaks caused by the sampling period are absent in the GLS periodogram.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.