2 Light curve observations

The just recent discovery of the eclipsing nature of CU Cnc is likely to be a consequence of its faintness but also, very importantly, of its shallow and short eclipses. The photometric observations published by D99 indicate a system with an orbital period of 2.77 days, eclipses of only $\sim$0.2 mag in depth and about 2 hours in duration. However, the high scatter of these light curves and their poor phase coverage prevented an accurate determination of the physical properties of CU Cnc.

We report here new photometry of CU Cnc in the Johnson R and I bands. Both the photometric accuracy (a few millimagnitudes) and the phase coverage (over 2100 observations) are sufficient to guarantee a reliable determination of the light curve parameters thus permitting a critical evaluation of stellar models. The observations were carried out with the Four College 0.8-m APT, which is equipped with a refrigerated Hamamutsu photoelectric detector and filters closely matching the standard Johnson system. Differential photometry was obtained in which HD 72093 (F8, V=7.80) was employed as the comparison star, while HD 72358 (F5, V=8.31) served as the check star. Although it is common practice to employ comparison stars that match the spectral type of the variable star, we preferred to avoid G, K or M stars because of concerns with magnetic-related intrinsic variability. In contrast, late F stars are known to be photometrically very stable. Note that only precise R and I photometry could be obtained with our instrumental setup because of the faintness of CU Cnc at shorter wavelengths. The system is so red that at the I band, its brightness rises up to about $I\approx9$ mag. No evidence of significant light variations (down to the few milimagnitude level) was found for the comparison-check star sets. The observations were reduced using photometric reduction programs at Villanova University (USA). Differential extinction corrections were applied, although these corrections were typically very small.

There is one feature that makes our photometric dataset especially noteworthy: the $\sim$2100 measurements were obtained over a period of only 110 days (from December 5, 1998 through March 25, 1999). Classified as a flare star, CU Cnc is expected to be magnetically active and display some level of brightness variability due to surface inhomogeneities. Thus, the photometric observations were acquired over a short time span to minimise possible variations in the out-of-eclipse shape caused by starspot migration. Also, special care was taken to schedule the observations near the eclipses so that a very dense phase coverage could be achieved. To compute orbital phases, we adopted the ephemeris of D99 but correcting the spectroscopic reference epoch to the primary eclipse:

$$\begin{array}{rcrcrc} T(\mbox{Min I})&=&{\rm HJD}~2450208.5068&+&2.771468 & E. \\ \end{array}$$

The phased light curves are presented in Fig. 1. Note the dense phase coverage, the relative shallowness of the eclipses and the out-of-eclipse variations presumably due to starspots. The individual photometric observations in the R (n=2157) and I (n=2188) bands are provided in Tables 1 and 2 (HJD, phase and $\Delta$mag), respectively, which are only available in electronic form.

Figure 1: R and I light curves of CU Cnc. A zero point shift of -1.45 has been applied to the R photometry for display purposes. The solid line in the bottom panel is the best-fitting synthetic light curve generated by W-D. The top panels show the residuals of the fits. Enlargements of the fits to the eclipse phases are shown in Fig. 3.

As mentioned above, CU Cnc has an optical companion at $\sim$12 $^{\prime\prime}$. The diaphragm used for the photometric observations was 45 $^{\prime\prime}$ in diameter, so the light of the companion was included in the measurements. To estimate exactly the fraction of contributed light we made use of CCD images taken with the ALFOSC instrument at the Nordic Optical Telescope (La Palma). Several images in the V and I passbands were kindly acquired by R. J. Irgens on May 10, 1999 (HJD 2451309.375) at an orbital phase of 0.215. The reduction was carried out in the classical manner and aperture photometry was performed to estimate the differential magnitude between CU Cnc and its companion. We obtained values of $\Delta V=1.45\pm0.01$ and $\Delta I=1.20\pm0.01$, which translates into the following fractions of 3rd light, that can be directly incorporated into the analysis of the light curves:

$$F_3^{V} = 0.21\pm0.01; \hspace{3mm} F_3^{I} = 0.25\pm0.01.$$

From these we estimate the fraction of 3rd light in the R passband to be $F_3^{R} = 0.23\pm0.01$.