DETECTION OF A GIANT EXTRASOLAR PLANET ORBITING THE ECLIPSING POLAR DP LEO

DP Leo is one of the AM Her-type stars, a group of strongly magnetic cataclysmic binaries also referred to as polars because of their strongly polarized optical emission. After its discovery as the first eclipsing polar more than 20 years ago as the optical counterpart of the EINSTEIN source E1114+182 (Biermann et al. 1985), DP Leo was continuously observed from the ground and the space with, e.g., the Hubble Space Telescope (HST; Stockman et al. 1994), ROSAT (Robinson & Cordova 1994), and XMM-Newton (Schwope et al. 2002; Pandel et al. 2002). It was discovered to be a two-pole accretor (Cropper & Wickramasinghe 1993) that shows asynchronous rotation of the white dwarf in the system (Robinson & Cordova 1994). However, with an extra spin of the white dwarf in the binary of about 2°–25 per year, the degree of asynchronism is seemingly much smaller than in the other four objects (Campbell & Schwope 1999). Absolute parameters of DP Leo were determined by several authors (e.g., Bailey et al. 1993; Schwope et al. 2002).

Epochs and orbital periods of DP Leo were determined by some investigators (e.g., Biermann et al. 1985; Schaaf et al. 1987; Stockman et al. 1994; Robinson & Cordova 1994; Schwope et al. 2002; Pandel et al. 2002) by using those mid-eclipse times derived with the X-ray, UV, and optical data. Biermann et al. (1985) pointed out that the data limit any period change to less than about dP/dt = 1.6 × 10⁻¹¹ s/s. Later, the long-term decrease of the orbital period was reported by Stockman et al. (1994), Robinson & Cordova (1994), Schwope et al. (2002), and Pandel et al. (2002). More recently, an orbital period investigation on the other eclipsing polar HU Aqr suggested that its period change is very complex and may show a cyclic variation or a combination of a long-term decrease and a cyclic change (Schwarz et al. 2009). To understand the properties of the period variation of DP Leo, we monitored it with the 2.4 m telescope in Lijiang station of the Yunnan Astronomical Observatory since 2009 March. Here we report the discovery of a cyclic change in the orbital period of DP Leo that we interpret as the presence of a tertiary component, most likely a giant extrasolar planet companion to this polar system.

The eclipsing polar DP Leo was monitored since 2009 March 3 by using the 2.4 m telescope in Lijiang station of the Yunnan Astronomical Observatory. During the observation, a VersArray 1300B CCD camera attached to the telescope, and no filters were used. The integration time for each CCD image is 20 s. PHOT (measure magnitudes for a list of stars) of the aperture photometry package of IRAF was used to reduce the observed images. The light curves observed on 2009 June 3 are displayed in Figure 1. By using our observations, five mid-eclipse times were obtained that were measured by fitting three straight lines to the data around ingress (or egress) with the least-squares method. This method is the same as that used for NN Ser (Qian et al. 2009a) and QS Vir (O'Donoghue et al. 2003; Qian et al. 2009b). The errors listed in Table 1 are the standard deviation values.

Zoom In Zoom Out Reset image size

All the available mid-eclipse times were compiled by Schwope et al. (2002) and have been converted to BJED. By using the linear ephemeris given by Schwope et al. (2002),

$$\begin{equation} {\rm Min.}\ I = {\rm BJED}\,2448773.215071+0.06236283691\times {E}, \end{equation} \tag{ 1 }$$

where BJED 2448773.215071 is the initial epoch and 0.06236283691 days is the orbital period, the (O − C)₁ values were calculated. The corresponding (O − C)₁ diagram is shown in Figure 1 along with the epoch number E. During the period analysis, mid-eclipse times with errors larger than 60 s were not used because these data usually show large scatters, and will mislead us to understand the general trend of the (O − C)₁ variation. As displayed in Figure 1, our data do not follow the general trend of the long-term period decrease proposed by the previous authors, e.g., Stockman et al. (1994), Robinson & Cordova (1994), Schwope et al. (2002), and Pandel et al. (2002). This suggests that the period change of DP Leo may be more complex than a simple decrease.

The (O − C)₁ curve plotted in Figure 1 indicates that the orbital period of DP Leo needs to be revised, and it appears that there is a cyclic variation as well. To describe the general (O − C)₁ trend satisfactorily, a cyclic variation is required with an additional revision on linear ephemeris (dashed line in Figure 1). By using the least-squares method, we determined

$$\begin{eqnarray} {\rm Min.}\ I&=&2448773.214896(\pm 0.000017)\nonumber \\ && +0.06236284126(\pm 0.00000000024)\times {E}\nonumber \\ & &+0.000365(\pm 0.000020)\sin [0\mbox{$.\!\!^\circ $}002578(\pm 0.000023)\nonumber\\ &&\times {E}+142\mbox{$.\!\!^\circ $}3(\pm 2\mbox{$.\!\!^\circ $}1)]. \end{eqnarray} \tag{ 2 }$$

The derived orbital period in this equation is slightly longer than that determined by Schwope et al. (2002). The cyclic oscillation has an amplitude of 31.5 s and a period of 23.8 years. The (O − C)₂ values calculated with the new linear ephemeris in Equation (2) are plotted in the middle panel of Figure 2 where the cyclic change is seen more clearly. After the periodic change was subtracted from the (O − C)₂ curve, the residuals are displayed in the lower panel where no variations can be found indicating that Equation (2) can give a good fit to the (O − C)₁ curve.

Zoom In Zoom Out Reset image size

The cyclic period changes observed in close binary stars containing at least one cool component star could be explained as the magnetic activity cycles of the cool star (i.e., the Applegate mechanism; Applegate 1992). In this mechanism, it is assumed that a certain amount of angular momentum is periodically exchanged between the inner and the outer parts of the convection zone, and therefore the rotational oblateness and thus the orbital period will vary when the cool component goes through its activity cycles. As in the cases of, HW Vir (Qian et al. 2008a), HS 0705+6700 (Qian et al. 2009c), NN Ser (Brinkworth et al. 2006; Qian et al. 2009a), HU Aqr (Schwarz et al. 2009), and Z Cha (Dai et al. 2009), the fully convective secondary component in DP Leo rotates mainly as a rigid body, and lacks the thin interface layer between a radiative core and a convective envelope, where dynamo processes are thought to concentrate at for solar-type stars (e.g., Barnes et al. 2005). The investigations of HW Vir, NN Ser, HU Aqr, and Z Cha, indicated that, to interpret the observed cyclic period changes with this mechanism, the required energies are much larger than the total radiant energy of the fully convective components, which suggests that the mechanism of Applegate cannot interpret the cyclic period variations of these binary systems. Moreover, as discussed by Qian et al. (2008a, 2008b), a more general explanation of the cyclic period changes in close binaries is the light-travel time effect via the presence of a third body.

Evidence has shown that the accretion rate of polars was variable (e.g., Schwarz et al. 2009). The mass transfer from the secondary to the white dwarf primary in DP Leo should cause the period increasing continuously by considering a conservative angular momentum. The changing accretion rate should result in the variation of the rate of the period increase and cannot explain the observed cyclic period change. As displayed in Figure 2, no long-term period increase was found, which may indicate that the accretion rate in DP Leo is too low to be observed at present. On the other hand, when the accretion geometry changes from the high to intermediate states, the accretion spot longitude variations should influence the observed mid-eclipse times. However, the estimation by Schwarz et al. (2009) for the other eclipsing polar HU Aqr indicated that the shift of the O − C values caused by the movement of the accretion spot is only about 2 s, which is much smaller than the amplitude of the (O − C)₂ oscillation (31.5 s). The observed (O − C)₂ oscillation of DP Leo is not caused by the accretion spot longitude variations. Finally, a significant fraction of AM Her systems have small degrees of asynchronism, which means that the cyclic change in the (O − C)₂ curve (see Figure 2) may be caused by the exchange of the rotational angular momentum of the white dwarf and the orbital angular momentum. However, the investigation by Schwope et al. (2002) suggests that, compared with the other systems, the degree of asynchronism of DP leo is much smaller, i.e., (P_orb − P_ort)/P_orb ∼ 10⁻⁶. Moreover, if this explanation is true, it is difficult to understand why the angular momentum exchange is cyclic with a small period of 23.8 years, and what is the driven physical mechanism?

Therefore, we analyzed DP Leo for the light-time effect that arises from the gravitational influence of a third companion. When the eclipsing binary orbits the barycenter of the triple system, the change of the relative distance from the Earth can result in the observed cyclic change in the O − C diagram. The same method has been used to detect companions in orbits around pulsating white dwarf stars (e.g., Kepler et al. 1991; Winget et al. 2003; Mullally et al. 2008). Since the sine fit seems quite good, we assumed the orbit of the third body to be circular. With the absolute parameters determined by Schwope et al. (2002), we derived the mass function and the mass of the tertiary companion as f(m) = 4.45(±0.73) × 10⁻⁷ M_☉ and M₃sin i' = 0.00600(±0.00055) M_☉, respectively. The relation between the mass M₃ and the orbital inclination i' is displayed in Figure 3. If the orbital inclination of the third body is larger than 2557, the mass of the tertiary component corresponds to M₃ ⩽ 0.014 M_☉, and it should be an extrasolar planet. Therefore, with 71.6% probability, the third body is a giant extrasolar planet (by assuming a random distribution of orbital plane inclination).

Zoom In Zoom Out Reset image size

However, according to the theoretical investigations, circumbinary planets were expected, at least initially, to have a nearly coplanar orbit with the central binaries (e.g., Bonnell & Bate 1994). If this is true, i.e., i' = 795, the circumbinary component in DP Leo should be a giant extrasolar planet with a mass of 6.39 M_Jupiter at a distance of 8.6 astronomical units (AU) from the binary. One of the most interesting things that we have learned about extrasolar planet for last 17 years is that they can exist almost anywhere. However, to date, only one extrasolar planet was discovered to be a companion to the only known hibernating cataclysmic variable QS Vir (Qian et al. 2009b), and one was found in the orbit around the short-period white-dwarf–red-dwarf binary NN Ser (Qian et al. 2009a). Three white dwarfs are known to be wide companions to stars hosting extrasolar planets (Mayor et al. 2004; Lagrange et al. 2006; Mugrauer et al. 2007). The detection of a giant planet orbiting the first discovered eclipsing polar will provide us with more knowledge on the formation and evolution of circumbinary planets.

The separation between the Jupiter-like planet and the central eclipsing polar is 8.6 AU. By considering a distance of 400  PC (Schwope et al. 2002), this separation corresponds to 0.0215 arcsec. Therefore, it is difficult to detect the giant planet by direct imaging such as the Degenerate Objects around Degenerate Objects (DODO) survey (e.g., Hogan et al. 2007; Burleigh et al. 2008). To check the presence of the giant planet companion to the eclipsing polar, more precise mid-eclipse times are needed in the future.

This work is partly supported by Chinese Natural Science Foundation (No. 10973037, No. 10903026, and No. 10778718), the National Key Fundamental Research Project through grant 2007CB815406, the Yunnan Natural Science Foundation (No. 2008CD157), and by the Special Foundation of the President and the West Light Foundation of the Chinese Academy of Sciences. New CCD photometric observations of the system were obtained with the 2.4 m telescopes in Lijiang station of Yunnan Observatory.