28 And in 1996: variable but near the level of photometric detectability

Rodríguez E., Rolland A., López-González M.J. and Costa V.
Instituto de Astrofísica de Andalucía, CSIC
P.O. Box 3004, E-18080 Granada, Spain

28 And is a bright low amplitude $\delta$ Sct type star with V=5.^m23, $\Delta$V=0.^m05 (full amplitude) and P=0.^d0693 (Rodríguez et al. 1994). Since this star was reported as a variable type $\delta$ Sct by Nishimura (1969), it has been photometrically observed by a number of authors. From these works it seems: 1) 28 And is a monoperiodic pulsator; 2) this star shows long term amplitude variations with a time scale of years (amplitude varying from 6.8($\pm$0.6) mmag in the year 1977 to 31.9($\pm$0.1) mmag in 1991, in the B band of the Johnson photometric system; 3) this star is a nonradial pulsator with l=2 (on the basis of the phase shifts and amplitude ratios between observed light and colour variations using Strömgren photometry).

Figure 7: Observed light curves of 28 And in the v band with the Fourier fitting versus Heliocentric Julian Day.

We have collected high quality simultaneous uvby photometric observations during September and October, 1996, using the 90 cm telescope at Sierra Nevada Observatory, Spain. In total, 1708 measurements were obtained for each of the four uvby filters on fifteen nights. As an example, the v observed light curves are plotted in Figure 7 with the Fourier fitting obtained below. This shows the amplitude of the light curve to be very small, less than 4 mmag ! (full amplitude, from peak to peak) in the filter v, being the filter with largest amplitude. When we average the b and v measurements to approximate the variations in B, a value of 1.72($\pm$0.07) mmag is derived for the main frequency. This value is very small when compared with any other data set found in the bibliography. In particular, a very strong decrease has taken place in the amplitude of the main pulsation between the years 1991 to 1996, being the latter amplitude about 19 times smaller than five years before. In addition to the main frequency $\nu$₁=14.4282 cd^-1 (in very good agreement with earlier determination by other authors) with an amplitude of 1.81 mmag in the filter v, the frequency analysis of the data also reveals the existence of a secondary frequency $\nu$₂=17.23 cd^-1 with very small amplitude (0.61 mmag in the same filter). Hence, the monoperiodicity is not confirmed using the new data. Figure 7 shows the spectral window and power spectra of 28 And in the v filter before and after (second and third panel) prewhitening for $\nu$₁. The last panel shows the power spectrum after removing $\nu$₁ and $\nu$₂. The line in the last two panels of Figure 7 indicates the significance level as described by Breger et al. (1996). After subtracting for the main frequency, the periodograms show some other peaks at about 4.63, 10.11 and 17.23 cd^-1, but only the last one seems to be statistically significant with an amplitude signal-to-noise (S/N) ratio equal to 4.1. The mean noise level in the region from 12 to 22 cd^-1 is 0.15 mmag. This peak was also confirmed in the

Figure 8: Power spectra of 28 And in the v filter before and after prewhitening the frequencies $\nu$₁ and $\nu$₂ to our data.

other two b and y filters, but not in the u band (in this last filter, the noise level is too high), and the amplitudes seem to be consistent with a pulsation mode. The other two peaks above mentioned are not significant. In both cases, the amplitude S/N ratio is smaller than 4.0 (3.4 and 2.9, respectively) and their positions in the periodograms are not confirmed in any of the other filters. Thus, only the frequency of 17.23 cd^-1, corresponding to the last peak, was included into the final solution. The same analysis was performed on the other filters and the results were consistent with each other. It is possible to gain some insight into the pulsation modes identification by using the method of the phase shifts and amplitude ratios between observed light and colour variations as described in Garrido et al. (1990). This way, Rodríguez et al. (1993) suggest nonradial nature with l=2 for the main pulsation of 28 And. Now, using the new data, the Fourier solution shows again that the phase shifts for $\nu$₁ and v, b and y bands are in the same order as in Rodríguez et al. (1993), that is, the light maximum occurs later when the wavelength is longer. Thus, a nonradial nature is again suggested for $\nu$₁, with l=2 as the most probable value. In the case of $\nu$₂, the phase value in the filter v gives us some indication in favour of nonradial pulsation. However, the error bars are much larger than in the former case and the light maxima in the b and y filters seem to be occurring at the same time. Following this, a definitive conclusion can not be made at this moment for $\nu$₂.

References
Breger M. Handler G., Serkowitsch E. et al. 1996, A&A 309, 197 Garrido R., García-Lobo E., Rodríguez E. 1990, A&A 234, 262 Nishimura S. 1969, Ap&SS 3, 77 Rodríguez E., Rolland A., López de Coca P., Garrido R., Mendoza E.E. 1993, A&A 273, 473 Rodríguez E., López de Coca P., Rolland A., Garrido R., Costa V. 1994, A&AS 106, 21