3.1. RX J2022.6-3954
3.1.1. Orbital period and modulations
The V-filter lightcurves of 2022-39 show pronounced dips of 0.5 mag, recurring about every 78 min which proved to be the orbital period of the system. Our CCD photometry covers two full orbits during one night in July 1995 with a time resolution of 3 min and five additional orbits spread over a baseline of 10 nights in November 1995 (time resolution 40-60 sec). From these data we have obtained eight dip timings (cf. Table 3) which are suitably well spaced that a period search could be done and alias periods could be eliminated. Based on the combined July and November 1995 photometry, a -fit to the dip timings yields the following ephemeris for the optical dip:
The V-filter lightcurves folded on this period are shown for the individual nights in Fig. 3b-e. The average brightness is somewhat variable from night to night (V 18.4-19.0), probably due to fluctuations of the accretion rate. The shape of the lightcurves is roughly constant with maximum flux occurring around orbital phase 0.8 to 0.9. A second although generally lower maximum is observed at phase 0.2-0.3. The full amplitude of the orbital variation is 0.8 mag. The equivalent widths of the H and emission lines (Fig. 3f,g) obtained from our July and November 1995 spectroscopy vary by about 50% over the orbit. Maximum equivalent widths are observed around phase 0.9 for H and around phase 0.0 for He II.
As our low-resolution spectroscopy does not resolve individual components of the Balmer and emission lines we have determined radial velocities by fitting single Gauss functions to the line profiles. Eventhough the equivalent width variations are different for the Balmer and emission lines the corresponding radial velocities do not significantly differ. The radial velocities show a strong modulation with the photometric period. The radial velocity amplitude derived from a sinusoidal fit to the average Doppler shifts of H , H , and He II is (454 18) km/s (Fig. 3h).
The RASS X-ray data folded over the optical period (Fig. 3a) show a discontinuity of the 0.1-2.4 keV flux during the X-ray bright phase when the flux decreases from 1 cts/s to the detection limit. Our optical ephemeris is sufficiently accurate to determine the orbital phase of the X-ray dip to be 0.92 0.10. It is not uncommon in AM Herculis systems that X-ray and optical dips coincide in orbital phase as both may be caused by an eclipse of the accretion region on the white dwarf by the accretion stream. If we include the timing of the X-ray dip in our period search, we find an improved dip ephemeris
3.1.2. Magnetic field strength
The optical spectra of 2022-39 show strong Balmer, He I, and He II emission lines with variable profiles characteristic of AM Herculis binaries. The underlying continuum is flat during the optical dip phase ( -0.05 - 0.05) and shows broad humps during the bright phase ( 0.15 - 0.30 and 0.75 - 0.92). These humps are more obvious in the difference spectrum between bright and dip phase spectra and are typical for cyclotron emission radiated from an accretion plasma in a strong magnetic field (cf. Fig. 4). We have identified the hump pattern in the difference spectrum with the 2nd, 3rd, and 4th cyclotron harmonics corresponding to a magnetic field strength B = (67 2) MG.
3.1.3. Distance estimate
In the dip phase optical spectra of 2022-39 no features related to the M dwarf secondary can be seen down to the noise level. This implies that the contribution of the secondary star to the total flux in the R band must be less than about 10%. Therefore, the method described by Bailey (1981 ) can be adapted to derive a lower limit for the distance of 2022-39. A Roche-lobe filling red dwarf with an orbital period of 1.3 hours is likely to have a radius of 0.13 and a mass of 0.1 (depending only weakly on ).
A single M-dwarf with similar mass has a spectral type later than M5 such as VB 8 which has infrared colours V-K = 7.92 and R-K = 5.83 (Reid & Gilmore 1984 ). Using the improved surface brightness relationship for M dwarfs (Ramseyer 1994 ) and V-K of VB 8, we obtain a K surface brightness =5.44 for the secondary in 2022-39. With R-K of VB 8 and R 22.1 from the Nov. 1995 dip phase spectrum of 2022-39, this leads to a lower limit pc for the distance of 2022-39.
3.2. RX J0132.7-6554
3.2.1. Orbital period and modulations
Due to the low mean brightness (V 20 mag) of the system, photometry of 0132-65 was obtained in white light in order to reach sufficient time resolution and signal-to-noise of the data. Time resolutions of 4 min and 1.5 min could be achieved with the 0.9-m and 2.2-m telescopes, respectively. The lightcurves obtained during both runs have similar shapes and modulation amplitudes. The mean amplitudes of the lightcurves are 1 mag. A bright phase with a symmetrical rise and decline lasts about 65% of the time and repeats about every 78 min. During the faint phase the shapes of the lightcurves are flat with some fluctuations. Seven optical maxima have been observed in two subsequent nights in September 1993 and further seven maxima distributed over 10 nights in November 1995 (cf. Table 3). A time series analysis of both data sets yields an optical period min.
Unfortunately, the baseline between both observing runs is too long to derive an unambiguous period from the combined data set. The optical lightcurves of the two runs folded on the best period are shown in Fig. 6 together with the ROSAT PSPC data obtained in November 1993. The latter are the only X-ray data of 0132-65 which cover an entire orbit with a sufficient countrate to generate a decent lightcurve. Our optical ephemeris is sufficient to get the phasing between the September 1993 photometry and the November 1993 X-ray data with an accuracy of . The maximum of the 0.1-2.4 keV soft X-ray flux occurs during the optically faint phase. As only 10 source counts have been obtained during our ROSAT HRI observation of 0132-65 no light curve could be produced from these data.
3.2.2. Magnetic field strength
The identification spectrum of 0132-65 (cf. Fig. 5) obtained in January 1992 clearly shows the typical optical emission lines (strong Balmer, He I, and He II) found in polars during their high-state. These lines are much weaker in our spectra obtained in July and Nov. 1995 (cf. Fig. 7) which show the system in a low state. In July 1995 only two spectra with exposure times of 12 min each were obtained as the weather was not very collaborative and the system turned out to be too faint to obtain time resolved spectroscopy. Even though the signal-to-noise ratio is not good, the difference of both spectra clearly shows humps in the optical continuum which we have identified as the 2nd, 3rd, and 4th cyclotron harmonics corresponding to a magnetic field strength of (68 2) MG (cf. Fig. 7, bottom). After improving the ephemeris using CCD photometry we obtained two spectra of 0132-65 in November 1995, one 40 min spectrum centered on the bright phase and one 30 min spectrum centered on the faint phase. The difference of these two spectra shows cyclotron humps at the same positions confirming the 68 MG magnetic field strength derived from the July 1995 spectra (cf. Fig. 7).
3.2.3. Distance estimate
Like in 2022-39 no features related to the M dwarf secondary can be seen in the faint phase optical spectra of 0132-65. Therefore, the contribution of the secondary star to the total flux in R must be less than 10%.
Using the same method as described in section 3.1.3for 2022-39 and R 23.1 from the Nov. 1995 faint phase spectrum of 0132-65 we obtain a lower limit pc for the distance of 0132-65.
3.3. X-ray spectra
The photon event files of the different datasets of 2022-39 and 0132-65 were binned in such a way that the signal to noise ratio is 2 in each energy bin. Like for other AM Her systems we have fitted two component (blackbody + thermal bremsstrahlung) models with different parameters to the data. The X-ray spectra are shown in Fig. 8. A summary of the fits we have performed to the data is given in Tab. 4. Due to the low number of photons detected and the poor spectral resolution of the ROSAT PSPC and k are poorly constrained for both objects. In order to somewhat restrict the parameters to physically meaningful values we included additional conditions and performed a series of fits to the spectra in which we held several parameters fixed. In all our fits we assumed that the contribution of the hard component in both systems can be represented by a k = 20 keV thermal bremsstrahlung model. The galactic column densities used below were obtained from Dickey and Lockman (1990 ) as implemented in the EXSAS software package.
Table 4. Summary of spectral fits to the survey and pointed PSPC data of 2022-39 (fits a-b) and 0132-65 (fits c-e). Parameters given without error range have been fixed for the particular fit. and are the unabsorbed fluxes in the 0.1-2.4 keV ROSAT band
For 2022-39 we fixed the column density to the galactic value and find that the temperature range for is 20-51 eV. The fluxes for both components can be taken from fits (a) and (b) in Tab. 4. Using the fluxes obtained in fit (a) with =25 eV we get the following ratio of bremsstrahlung vs. blackbody fluxes in the ROSAT band, .
For fits (c) and (e) to the 0132-65 data we used obtained in fit (d) which is about 2/3 of the galactic as these are less well constrained than fit (d). leads to a 2- temperature range = 15-48 eV for the blackbody component of 0132-65 . Even though the system was a factor 5 brighter during the RASS than during the pointed observations. The flux ratio 0.021 remains almost the same. Both low values of fit nicely into the picture presented by Beuermann (1997 ) in which the bremsstrahlung component is gradually suppressed as the magnetic field strength increases.
© European Southern Observatory (ESO) 1997
Online publication: April 8, 1998