Robert Andrew Donahue

No planet for HD 166435

The G0 V star HD 166435 has been observed by the ber-fed spectrograph ELODIE as one of the targets in the large extra-solar planet survey that we are conducting at the Observatory of Haute-Provence. We detected coherent, low-amplitude, radial-velocity variations with a period of 3.7987 days, suggesting a possible close-in planetary companion. Subsequently, we initiated a series of high-precision photometric observations to search for possible planetary transits and an additional series of Ca II H and K observations to measure the level of surface magnetic activity and to look for possible rotational modulation. Surprisingly, we found the star to be photometrically variable and magnetically active. A detailed study of the phase stability of the radial-velocity signal revealed that the radial-velocity variability remains coherent only for durations of about 30 days. Analysis of the time variation of the spectroscopic line proles using line bisectors revealed a correlation between radial velocity and line-bisector orientation. All of these observations, along with a one-quarter cycle phase shift between the photometric and the radial-velocity variations, are well explained by the presence of dark photospheric spots on HD 166435. We conclude that the radial-velocity variations are not due to gravitational interaction with an orbiting planet but, instead, originate from line-prole changes stemming from star spots on the surface of the star. The quasi-coherence of the radial-velocity signal over more than two years, which allowed a fair t with a binary model, makes the stability of this star unusual among other active stars. It suggests a stable magnetic eld orientation where spots are always generated at about the same location on the surface of the star.

ACTIVITY-RELATED RADIAL VELOCITY VARIATION IN COOL STARS

Planets have been detected orbiting several solar-type stars with the use of high-precision radial velocity (vr) measurements. While changes in vr can be measured with an accuracy of a few meters per second, there has been relatively little study of how other astrophysical processes, such as magnetic activity, may affect the observed velocities. In this paper, we use published data and simple models to explore the contributions to vr from two activity-related sources, starspots and convective inhomogeneities, as these features rotate across the disk and evolve in time. Radial velocity perturbations due to both of these sources increase with rotation and the level of surface activity. Our models indicate that for solar-age G stars, the amplitude of perturbations due to spots is AS 5 m s-1, increasing to AS ~ 30-50 m s-1 for Hyades-age G stars. If fS is the starspot area coverage, we find that A -->Sf -->0.9S v sin i. The effects of convective inhomogeneities (as observed in line bisector variations) appear to depend on both rotation and spectral type. Young (active) F and G dwarfs can have convective vr perturbations with amplitudes AC 50 m s-1, while vr amplitudes are reduced for stars with lower v sin i and cooler Teff. We show that vr data from the literature display similar trends with v sin i and Teff. AS and AC will be strongest at or near timescales related to magnetic activity variations: rotation, active region growth and decay, and activity cycles. Thus, knowledge of these timescales and typical AS and AC values are important in searching for extrasolar planets, especially those around younger, more active stars or those with small vr reflex amplitudes (i.e., 20 m s-1). We discuss implications of our results for current planet detections and planet search strategies.

Patterns of Photometric and Chromospheric Variation among Sun-like Stars: A 20 Year Perspective

We examine patterns of variation of 32 primarily main-sequence Sun-like stars [selected at project onset as stars on or near the main sequence and color index 0.42 ≤ (B - V) ≤ 1.4], extending our previous 7-12 yr time series to 13-20 yr by combining Stromgren b, y photometry from Lowell Observatory with similar data from Fairborn Observatory. Parallel chromospheric Ca II H and K emission data from the Mount Wilson Observatory span the entire interval. The extended data strengthen the relationship between chromospheric and brightness variability at visible wavelengths derived previously. We show that the full range of photometric variation has probably now been observed for a majority of the program stars. Twenty-seven stars are deemed variable according to an objective statistical criterion. On a year-to-year timescale, young active stars become fainter when their Ca II emission increases, while older less active stars such as the Sun become brighter when their Ca II emission increases. The Suns total irradiance variation, scaled to the b and y stellar filter photometry, still appears to be somewhat smaller than stars in our limited sample with similar mean chromospheric activity, but we now regard this discrepancy as probably due mainly to our limited stellar sample.

Photometric and Ca II H and K Spectroscopic Variations in Nearby Sun-like Stars with Planets. III.

We present the results of an analysis of time-series photometry, Ca II H and K spectrophotometry, and high-dispersion visible spectra of nine nearby Sun-like stars recently identified as having planets. For the six stars whose presumed planets have orbital periods of less than 4 months (? Boo, 51 Peg, ? And, ?1 Cnc, ? CrB, and 70 Vir), sine-curve fits to the photometric data show no variations with semiamplitude greater than 1 or 2 parts in 104. Photometric variations in 47 UMa are similarly small, although our photometric data of this star are slightly affected by variability of the comparison star. Nonvariability at this level of precision is sufficient to rule out surface magnetic activity as the cause of the observed radial-velocity variations in these seven stars and makes nonradial pulsations unlikely as well. Thus, our photometry provides indirect but strong support for true reflex motions?planets?in these seven stars, but cannot yet so support the planetary hypothesis for the two additional stars, 16 Cyg B and Gl 411. Continued photometric monitoring of the short-period systems may soon result in the direct detection of these planets in reflected light. We have used our photometric fluxes to search for possible transits of the extrasolar planets. Transits definitely do not occur in ? Boo, 51 Peg, ? And, and ?1 Cnc, and probably do not occur in ? CrB and 70 Vir. Our transit-search results are inconclusive for 47 UMa, and we cannot address the issue for 16 Cyg B and Gl 411. The precision of our photometry is sufficient to detect transits of planets even if they are not gas giants, as currently assumed, but much smaller objects with rocky compositions. The chance of finding at least one transit in the six stars is ~40%. We find significant year-to-year photometric variability only in ? Boo, which is not only the youngest star in the sample but also the star with the shallowest convective zone. The interseasonal range in its yearly mean photometric flux is ~0.002 mag, roughly twice the 0.0008 mag decadal variation in the Suns total irradiance. Monitoring of the relative Ca II H and K fluxes began between 1966 and 1968 for 51 Peg, ? Boo, ? CrB, and Gl 411, between 1990 and 1993 for 47 UMa, 70 Vir, 16 Cyg B, and ?1 Cnc, and in 1996 for ? And. The data have been newly recalibrated for improved long-term instrumental stability, resulting in better precision of the Ca II records. Five of the nine stars in this study have little or no detectable year-to-year variation in Ca II flux. The remaining four show moderate or pronounced variability: ? Boo, whose radial-velocity and photometric variations have comparatively high amplitudes; Gl 411, whose planetary companion was inferred astrometrically, not spectroscopically; ?1 Cnc, which may undergo decadal cyclic activity; and ? And, which shows moderate year-to-year variability. Except for 47 UMa, intraseasonal variability consistent with rotation was detected in the Ca II records of all stars. However, the rotation periods determined for ? And, 70 Vir, and 16 Cyg B are of low confidence. An examination of the recalibrated Ca II records for 51 Peg finds a rotation period of 22 days, in contrast to our previous result of 37 days. Ages have been estimated from the mean Ca II flux and, where possible, the rotation period. We find general consistency with the ages determined by others comparing properties determined from high-resolution spectroscopy to evolutionary models, although the uncertainties are, in general, large.

Properties of Sun-like Stars with Planets: ρ1 Cancri, τ Bootis, and υ Andromedae

Planets have been reported orbiting the Sun-like stars ρ1 Cnc, τ Boo, and υ And based on low-amplitude radial velocity variations. We have derived information on the first two stars from analysis of spectra, as well as parallel records of high-precision Stromgren b and y photometry and Ca II H + K fluxes. In the case of ρ1 Cnc, the upper limit (peak to peak) of nondetection of photometric variability at the orbital period is Δy ~ 0.0004 mag. The possibility of a planetary transit cannot be ruled out completely from the photometric data. Variations of the Ca II fluxes suggest a rotational period of ~42 days, in agreement with the inferred v sin i ~ 2 km s-1. The age of ρ1 Cnc is ~5 Gyr, based on its average Ca II flux and a relation between Ca II flux and age. The star τ Boo, unlike the other reported solar-type stars with planets, is relatively young (~2 Gyr). Despite its young age, it is photometrically nonvariable at the orbital period with an amplitude of Δ(b + y)/2 ~ 0.0004 mag (peak to peak); however, small-amplitude interseasonal variability is seen. No planetary transits were found in the photometry, which limits the inclination of the planets orbital plane to Earths line of sight to less than 83° (where 90° is coplanar). The Ca II record shows a weakly significant rotational period near 3.3 days, coincident with the orbital period of the companion. The Ca II record also shows a period of 116 days that has persisted for 30 years and is not seen in the photometric record. The persistence and timescale of this Ca II variation mean that it has no counterpart in Sun-like magnetic activity. The amplitude of the reflex velocity of the parent star (~450 m s-1) is much larger than the radial velocity perturbations expected from the presence of either surface inhomogeneities or line-bisector variations. Thus the anticipated perturbations from those stellar effects do not refute the inference of reflex velocities. We have few Ca II flux measurements for υ And. Its age and rotational period are estimated to be ~5 Gyr and 12 days, respectively. Our results for ρ1 Cnc and τ Boo are consistent with the explanation of planets as the cause of the velocity variations.

PRECISE AUTOMATIC DIFFERENTIAL STELLAR PHOTOMETRY

The factors limiting the precision of differential stellar photometry are reviewed. Errors due to variable atmospheric extinction can be reduced to below 0.001 mag at good sites by utilizing the speed of robotic telescopes. Existing photometric systems produce aliasing errors, which are several millimagnitudes in general but may be reduced to about a millimagnitude in special circumstances. Conventional differential photometry neglects several other important effects, which are discussed in detail. If all of these are properly handled, it appears possible to do differential photometry of variable stars with an overall precision of 0.001 mag with ground based robotic telescopes.

A False Planet around HD 192263

We present new high-precision Stromgren photometry and Ca II H and K spectrophotometry of HD 192263. Based on radial velocity variations detected previously by two groups, this K2 V star was thought to host a 0.75 MJup (minimum mass) planetary companion in a 24 day orbit. Our photometric observations reveal periodic variations that match the purported planetary orbital period, while the Ca II H and K emission fluxes are modulated on half the planetary period. This suggests that rotational modulation of the visibility of stellar surface activity is the source of the observed radial velocity variations. Therefore, HD 192263 should be removed from lists of stars with well-established planetary companions unless further observations and analysis can support the existence of the planet in spite of the stars intrinsic variations.

A STUDY OF VARIABILITY IN A SAMPLE OF G AND K GIANTS

Eight years of Ca II surface activity records from Mount Wilson Observatory measured for 12 bright G-K III stars have been analyzed in order to detect periodic variations attributable to rotation. We also present photometric V-band data for these stars from the Fairborn 0.25m Automatic Photometric Telescope (APT) that yielded a photometric period in one case and rms deviations from apparently constant brightness levels for the remaining 11 stars. The Ca II data yielded rotation periods for 10 out of 12 giant stars. We demonstrate that the photometric variability and non-variability of these stars can be predicted from their Rossby numbers calculated from our observed rotation periods and convective turnover times scaled up from the main sequence.

The solar surface differential rotation from disk-integrated chromospheric fluxes

Disk-integrated solar chromospheric Caii K-line (3933.68 Å) fluxes have been measured almost daily at Sacramento Peak Observatory since 1977. Using observing windows selected to mimic seasonal windows for chromospheric measurements of lower Main-Sequence stars such as those observed by Mount Wilson Observatorys HK Project, we have measured the solar rotation from the modulation of the Caii K-line flux. We track the change of rotation period from the decline of cycle 21 through the maximum of cycle 22. This variation in rotation period is shown to behave as expected from the migration of active regions in latitude according to Maunders ‘butterfly diagram’, including an abrupt change in rotation period at the transition from cycle 21 to cycle 22. These results indicate the successful detection of solar surface differential rotation from disk-integrated observations. We argue that the success of our study compared to previous investigations of the solar surface differential rotation from disk-integrated fluxes lies primarily with the choice of the length of the time-series window. Our selection of 200 days is shorter than in previous studies whose windows are typically on the order of one year. The 200-day window is long enough to permit an accurate determination of the rotation period, yet short enough to avoid complications arising from active region evolution. Thus, measurements of the variation of rotation period in lower Main-Sequence stars, especially those that appear to be correlated with long-term changes in chromospheric activity (i.e., cycles), are probably evidence for stellar surface differential rotation.

Evidence of differential surface rotation in the solar-type star HD 114710

Observations of the chromospheric Ca II H and K emission variability of the intermediate-age solar-type star HD 114710 (β Comae Berenices, G0 V) obtained at Mount Wilson Observatory over the past 10 years reveal a secular change in the seasonal rotation period that can be interpreted as surface differential rotation. The dependence of rotation period on chromospheric flux (i.e., activity-cycle phase) suggests that the star may have two latitudinal zones of activity: one in which changes in rotation period appear to follow the starspot activity cycle, and another confined to a narrow range of periods that does not