next previous
Up: Photometric variability of

2. Observations and data reduction

CCD photometry was carried out from 1993 to 1995 at the 0.6 m telescopes of OPD with a front illuminated GEC P8603A CCD detector with resolution of about tex2html_wrap_inline1460 (385tex2html_wrap_inline1462 578  pixels). No attempt was made to transform the computed magnitudes to a standard system since our main interest is to study time variations. This procedure does not require photometric sky conditions and successfully overcomes the problematic effects of variable transparency. In deciding between observing a large sample of diverse objects or fewer objects at more colors, we chose the former. Therefore, our images were taken with the V and I filters, and exposures were all 3 minutes long.

Photoelectric differential photometry in tex2html_wrap_inline1468 were also taken with the same telescopes using the FOTRAP (Jablonski et al. 1994). This sample includes recently identified TTS sorted by their IRAS colors and conspicuous Li lines (Pico dos Dias Survey - henceforth PDS; Gregorio-Hetem et al. 1992, henceforth GHETAL92). Special attention is given to AS 216 and AS 218, monitored since 1985 to establish the stability of their rotational periods.

In general the objects are brighter than tex2html_wrap_inline1470 and cooler than K5. T Cha (G8) is included because of its reportedly extreme variability.

2.1. CCD photometry

We adopt the standard method to perform the differential photometry. The stars in each field - excluding the variable target - are examined for stability and those showing no trace of variability are chosen as comparisons. The uncertainty in the photometry is estimated from the dispersion in the magnitude differences and ranges from 0.01 to 0.02 mag. The TTS observed with the CCD are SY Cha, TW Cha, CT Cha, VZ Cha, WY Cha, SZ 45, T Cha, SZ 77, SZ 82 and SZ 108.

Data reduction is carried out with the Image Reduction and Analysis Facility (IRAF) software. Images are corrected for electronic bias, pixel-to-pixel gain variations (with dome flats), and cosmic rays are removed. The routines of the APPHOT package are used to obtain relative magnitudes for the objects selected in each field. The sky contribution is the modal value inside a specified annulus whose internal radius and width are automatically determined for each image by the RADPROF routine. This procedure avoids possible effects of nearby stars or defective pixels. Typical values of the aperture range from four to seven pixels, corresponding to 2.3'' to 4'' respectively.

2.2. Photoelectric photometry

Photoelectric photometry was performed using the FOTRAP (Jablonski et al. 1994). This device has a rotating filter wheel (1200 rpm) that enables measurements in 5 colors in a quasi-simultaneous way. The photometric system is matched to the Johnson/Cousins tex2html_wrap_inline1468 standard system. We achieve differential magnitudes by observing the stars in the order: TTS, comparison 1, comparison 2 and TTS. The two arrays of differential values per season (TTS - Comp) are then combined and used as input to the periodogram analysis. The final set of observed TTS include all the listed PDS objects and AS 216, AS 218, FK Ser.

Table 1 summarizes our observation logs. The first and second columns list the names of the objects and are followed by the year of the observation. Columns 4 through 8 list the tex2html_wrap_inline1478 of the differential magnitudes. The number of observations in each of the observed seasons are indicated next. The spectral types are taken from the Herbig & Bell catalogue (1988 - henceforth HBC) or from a spectral type classification method developed to utilize the medium resolution data of the PDS survey. This method essentially compares the unknown TTS spectrum to those of standards of different spectral class observed with the same instrument and spectral format. The spectral type of T Cha is taken from Alcalá et al. (1993). Complete electronic tables with the photoelectric magnitudes can be accessed at the CDS, via anonymous ftp to cdsarc.u-strasbg.fr or via http://cdsweb.u-strasbg.fr/Abstract.html.

   

Star Other name Yeartex2html_wrap_inline1482 tex2html_wrap_inline1484tex2html_wrap_inline1486tex2html_wrap_inline1488 tex2html_wrap_inline1490NPSpT
PDS 01 Hen 1 1989 0.01 0.02 0.03 0.03 0.02 4K1
45 CoD-29 8887 1989/90 0.15 0.06 0.04 0.04 0.04 10 M3
50 Hen 600 1989/90 0.17 0.05 0.02 0.02 0.02 13M3+M3
59 GSG 9419-1065 1991 0.34 0.11 0.11 0.10 0.07 9 M1
66 Hen 892 1990 0.12 0.05 0.03 0.03 0.02 10K2
70 CoD-40 8434 1990 0.06 0.04 0.03 0.03 0.02 9K7
77 CoD-41 10484 1990 0.47 0.43 0.37 0.33 0.29 7K0
81 GSC 6209-923 1989/90 0.05 0.05 0.04 0.04 0.03 11K0
82 VV Sco 1990 0.17 0.15 0.10 0.08 0.06 7K0 + M1
83 V896 Sco 1990 0.35 0.24 0.19 0.15 0.12 7K5
89 GSC 6217-126 1990 0.12 0.080.07 0.06 0.04 4M0
99 1989 0.31 0.19 0.16 0.14 0.13 4M0
101 BZ Sgr 1989 0.91 0.66 0.53 0.42 0.34 7K0
HBC 565 SY Cha 1994/ 1995 ---- 0.23--0.14 22 M0
567 TW Cha 1994/1995 -- --0.29--0.14 18 M0
570 CT Cha 1995 ----0.08--0.04 10 M0
578 VZ Cha 1994/1995 -- --0.46-- 0.23 15K6
583 WY Cha 1994/1995 -- -- 0.08--0.07 11K7
590 Sz 45 1994/1995 ---- 0.15-- 0.0510 M0.5
591 T Cha 1995 -- -- 0.90--0.677 G8
603 Sz 77 1994/1995 -- -- 0.13 --0.0810 M0
605 Sz 82 1995/1994-- -- 0.07--0.0516 M0
620 Sz 108 1994/1995 ----0.04 --0.06 10 M0
656 AS 216 1985 0.21 0.09 0.05 -- -- 21K2
1986 0.24 0.11 0.06 -- -- 67
1987 0.23 0.09 0.05 -- -- 41
19890.44 0.14 0.08 0.05 0.04 7
1990 0.45 0.16 0.09 0.07 0.06 11
657 AS 218 1986 0.93 0.30 0.21 56K7
1989 1.47 0.70 0.40 0.24 0.12 5
1990 0.49 0.13 0.08 0.07 0.07 11
663 FK Ser 1986 0.890.320.1830 K5/K7
1989/1990 0.60 0.33 0.31 0.31 0.24 11

Table 1: Summary of the observed Southern T Tauri Stars. The star names are listed in the first and second columns. The seasons are indicated in the third column. Columns 4 through 8 list the tex2html_wrap_inline1478 of the differential magnitudes. The number of observations in each season are listed in Col. 9. The last column contains spectral types

2.3. Periodogram analysis

For all the stars but Hen 1, V896 Sco and PDS 89, we gathered enough observations to search for photometric periodicity. Since the observations were generally taken near meridian transit, our frequency search is limited to the interval 0.0 to 0.5 day-1, the quasi-Nyquist frequency. We cannot discriminate these from the aliases which occur in the frequency interval between 0.5 and 1.0 day-1.

We use the two following methods to access frequencies:

1) The Date Compensated Discrete Fourier Transform (DCDFT, Ferraz-Mello 1981) that computes the correlation between a sinusoid-plus-constant curve with the photometric data. The significance of the candidate peak is assessed with the method outlined by Ferraz Mello & Quast (1987), and gives the probability that a peak in the distribution is not formed by chance (the acceptance of a given period). The method also gives the error bar of the computed period.

2) The method developed by Akerlof et al. (1994, henceforth BSP) that computes the least-square fit of a sum of cubic B-spline functions to the period-folded light curve. The chosen frequency is that which provides the minimum tex2html_wrap_inline1496 statistic. A characteristic of this method is that the computed frequency is not affected by any predefined light curve shape (e.g. sinusoidal). This degree of freedom allows a range of possible solutions for time series with less than 10 entries. We tested the efficiency of this program using the photometry of several TTS available in electronic format (Herbst et al. 1994). The program provides firm solutions for stars with large numbers of entries (though sometimes disagreeing with the published value) and handles data sets with fewer entries (down to 7), though not claiming a unique solution for those cases. The final range of solutions can be narrowed down to a few options if the shape of the folded light curve is fixed to a sinusoidal one.

Our basic procedure which leads to a final rotational period is as follows. First, we utilize the BSP code for the V, R and I colors and select the 15 frequencies, in each color, yielding light curves of low tex2html_wrap_inline1496. The final solution tends to be unique if the inputed time series has a large number of entries. In this case, that frequency is easily identified in every color and distinguishable from the others by its low tex2html_wrap_inline1496 (see example in Akerlof et al. 1994). In some cases, we have a modest data set for which, as discussed above, the BSP code ascribes several possible solutions. In these cases, we select as the final BSP solutions those where the shape of the U, B, V, R and I folded light curves are correlated in phase. In addition, the curve shall present an even distribution of data-points in phase. In some cases, we are satisfied with more than one final BSP solution. Secondly, we run the DCDFT code and search for the best frequency along the steps indicated by this method. We analyze individual color bands, selecting the solutions yielding good phase coverage, large acceptance in all colors (U excluded), and consistency among the period-folded light curves for each color band. The chosen DCDFT period is then compared with that of the BSP method and usually the sinusoidal-light curve solution of the former method is adopted.


next previous
Up: Photometric variability of

Copyright by the European Southern Observatory (ESO)