We present a brief discussion of solar-type stars, what is known about them, why they are ideal targets for precise radial-velocity observations, and some open questions.

We present the first results of a long-term study to increase the external accuracy of (relative and absolute) radial velocities of early-type stars below the 1 km/s level. The accuracy suffers mainly from important systematic errors caused by intrinsic object-template mismatch in cross-correlation. Based on synthetic spectra, the magnitude of these mismatch errors is shown in different spectral regions and for different rotational velocities around spectral-type mid-A. A line selection and combination scheme is proposed which optimizes the accuracy (systematic errors) in view of the precision (random errors) required.

Several different high-precision radial-velocity programs are now underway at The University of Texas. This paper discusses the aspects of these programs that are related to the problem of detection of extrasolar planetary systems. This includes the McDonald Observatory Planetary Search program on the McDonald 2.7-m Harlan Smith Telescope, an accompanying program of high-resolution stellar line profile measurement, the European Southern Observatory planetary search program, the Keck Hyades survey, and the Hobby∙Eberly Telescope planet surveys. Here, we summarize each of these programs, and present recent results from each.

We have measured the radial velocities of 540 G and K main sequence stars with a precision of 3−10 ms−1 using the Lick and Keck échelle spectrometers. We had detected 6 companions that have m sin i < 7 MJup. We announce here the discovery of a new planet around Gliese 876, found in our Doppler measurements from both Lick and Keck. This is the first planet found around an M dwarf, which indicates that planets occur around low-mass stars, in addition to solar-type stars. We combine our entire stellar sample with that of Mayor et al. to derive general properties of giant planets within a few AU of these stars. Less than 1% of G and K main sequence stars harbor brown dwarf companions with masses between 5 and 70 MJup. Including Gliese 876b, 8 companions exhibit m sin i < 5 MJup which constitute the best planet candidates to date. Apparently, 4% of stars have planetary companions within the range m sin i = 0.5 to 5 MJup. Planets are distinguished from brown dwarfs by the discontinuous jump in the mass function at 5 MJup. About 2/3 of the planets orbit within just 0.3 AU due in part to their favorable detectability, but also possibly due to a real “pile up” of planets near the star. Inward orbital migration after formation may explain this, but the mechanism to stop the migration remains unclear. Five of eight planets have orbital eccentricities greater than that of our Jupiter, eJup = 0.048, and tidal circularization may explain most of the circular orbits. Thus, eccentric orbits are common and may arise from gravitational interactions with other planets, stars, or the protoplanetary disk. The planet-bearing stars are systematically metal-rich, as is the Sun, compared to the solar neighborhood.

This paper discusses the distinction between extra-solar planets and low-mass secondaries, in principle as well as in practice. Adopting a distinction based on the presumed different processes of formation, the paper compares the characteristic features of the giant planets in our solar system with those of the low-mass secondaries in spectroscopic binaries. The discussion reveals that there is no a priori obvious feature that can identify planets. Instead, this work considers the extremely small emerging population of discovered extra-solar planets. Based on the nine “planet-candidates” discovered as of mid-1998, it was found that their mass distribution is remarkably different from the distribution of low-mass secondaries. The transition between the two populations probably occurs at 10−30 Jupiter masses. This transition could reflect the borderline between planet and brown dwarf secondary masses.

The Advanced Fiber-Optic Echelle (AFOE) is a fiber-fed bench-top spectrograph specifically designed for precise radial velocity observations. The AFOE is permanently located at the 1.5-m telescope at Smithsonian’s Whipple Observatory in Arizona and is regularly used for monitoring exo-planet candidate stars and for asteroseismology observations. In this paper, we discuss the status of the instrument, as well as an upgrade to the instrument, a Fabry-Perot reference, which may prove important both for the AFOE and for all precise radial velocity (PRV) facilities.

Results are presented from the first five years of the high precision RV survey carried out with the 1.4m CAT+CES spectrograph at ESO La Silla. This RV survey of 37 solar-type stars was begun in Nov. 1992. Using an iodine gas absorption cell for self-calibration we currently achieve a long-term precision of 20 m/s in a 30-min exposure of a 5.5 mag star. This value is the typical ‘working’ precision in survey work, i.e. an average over all observing conditions.

We present the RV data for our most variable star, ι Hor (G0V), with a possible periodicity of 600 d that could imply a companion of minimum mass m sin i = 2.0 MJup. However, these data are severely complicated by stellar activity to which a large part of the variability must be attributed. Our RV data for ϕ2 Pav (F8V) contain a possible period of 43 d which would imply a planet with m sin i = 0.69 MJup. It must be stressed that for both stars the significance of the periods is still insufficient to conclude that planets are indeed orbiting them. We also show our data for a Hipparcos astrometric brown dwarf candidate, GJ 570A, that do not confirm expectations for a correspondingly high RV signal. Longer monitoring of the star is required in order to confirm a possible low amplitude variation.

Continuing precise radial velocity observations of ρ Coronae Borealis have allowed the determination of updated parameters of the 40-day orbit of its Jupiter-mass companion. This confirms the previously reported period and amplitude, and shows a small but marginally significant non-zero eccentricity. It also provides improved predictions for the times of possible transit of the companion in front of the star. The new data provide upper limits to the mass of a possible second companion to the system. The orbital parameters are discussed in the light of scenarios for the origin and migration of extra-solar giant planets.

The rapidly oscillating Ap stars (roAp) are a class of nonradially pulsating stars oscillating in low-degree modes with periods of 4–15 minutes. We have started a program to study the oscillations on several roAp stars using precise radial velocity (RV) measurements. The typical mean RV amplitude for the roAp stars we have observed is 50–400 ms−1, but this amplitude depends on the spectral region used for the measurement of the RV amplitude. A detailed line-by-line analysis reveals that the pulsational amplitude depends not only on atomic species, but on the line strength as well. For a given atomic species weak spectral lines exhibit a pulsational amplitude 10–100 times higher than for strong lines. The elemental effect can be understood in the context of the inhomogeneous distribution of elements known to occur on these stars and that is believed to result from the global dipole magnetic fields that are present. For instance, if an element is concentrated near the magnetic pole then it may have a higher RV amplitude than one that is distributed about the magnetic equator. The line strength effect is interpreted as arising from vertical structure to the pulsations since weaker lines are formed, on average, deeper in the atmosphere than stronger lines. Precise RV measurements may prove to be a powerful tool for probing both the vertical and horizontal structure of the pulsations in roAp stars.

We have observed Procyon with the fiber-fed échelle spectrograph ELODIE coupled with a Fabry-Perot etalon at the 193-cm of the Observatoire de Haute Provence (France) over 5 nights during 1997 December – 1998 January. Here, we present the results of a search for solar-like oscillations on this star and the performance of this instrumentation for asteroseismology. The power spectra show an excess of signal between 0.42 mHz and 1.46 mHz, which could be due to stellar oscillations.

We present precise stellar radial velocity (RV) measurements of the rapidly oscillating Ap (roAp) star 33 Lib taken in rapid succession over a 3-hr time span. A Fourier analysis of the data clearly shows the 8.2 min. pulsation period found previously by photometric investigations and gives a peak-to-peak (2K) amplitude of about 80 ms−1. We find, like in other roAp stars we have studied, that the RV amplitude depends on the spectral region used for measuring the pulsational RV amplitude and is as high as 57 ± 4.7 ms−1 in the region 5411–5500 Å and as low as 7 ± 3 ms−1 in the 5877–5976 Å region. An analysis of individual spectral lines show considerable scatter in the RV amplitude, ranging as high as 320 ms−1 and as low as 7 m−1. There is an overall trend of increasing RV amplitude with decreasing line strength. We also found that spectral lines due to nickel have a higher mean RV amplitude than chromium lines. We believe that the line strength variations result from the vertical atmospheric structure of the pulsations and that the elemental differences are related to the inhomogeneous distribution of elements known to occur on Ap stars. Precise stellar radial velocity studies of roAp stars may be a powerful tool for studying both the spatial (surface) and vertical structure to the pulsational velocity field.

High accuracy measurements of variations in the radial velocity of the K1 giant star Arcturus have been obtained. The observations span 5 years and have a point-to-point repeatability of 5 ms−1 and night-to-night stability of better than 20 ms−1. Velocity oscillations of Arcturus were discovered during the course of this work in 1986. Subsequent, extensive additional data, indicate that Arcturus is exhibiting global nonradial acoustic oscillations with characteristics similar to those occurring in the Sun.

All observations were done using a radial velocity spectrometer, designed to search for extrasolar planets, at a dedicated facility of the University of Arizona on Kitt Peak. A dedicated facility was crucial to this work — because of the changing nature of the oscillations, many observing runs, over several years, were required to understand the star’s behavior. Continuous data sets as long as 30 days were acquired. Preliminary pulsation models were performed in collaboration of Art Cox at Los Alamos National Lab.

The velocity power spectra are complicated and variable. There is substantial evidence that the variations are solar-like p-mode oscillations. At least 10 frequencies have been identified, over the range 8.3 to 1.7 days. A spectrum of evenly spaced modes is apparent, yielding a value for Δν0 ≈ 1.2 μ Hz. The average power spectrum peaks near 3 days. There is a broad envelope of power with a distribution reminiscent of that seen in the Sun. Both the mode spacing and the period of peak power are consistent with scaling from the Sun (Kjeldsen & Bedding 1995). The oscillations appear to undergo abrupt discontinuities and have phase coherence times of a few weeks. We interpret the driving to be due to stochastic excitation by convection.

Recent observations of the G5 IV star η Boo by Brown et al. (1997) have failed to confirm the detection of p-mode oscillations reported by Kjeldsen et al. (1995) Thus, Arcturus may be one of the first stars known to exhibit solar-like oscillations. If other K-giant variables can be shown to exhibit similar oscillations, Arcturus may represent the prototype for a new class of variable stars. We know some other K-giants are variable on short time scales (Hatzes and Cochran 1994b; Edmonds & Gilliland 1996), but many are not (Horner 1996).

Nearly a decade ago, “yellow giants” were introduced as a new class of low-amplitude radial-velocity variable stars. In this report we discuss new results for 12 spectral type K and early-M giants based on long-term monitoring using both the hydrogen-fluoride and iodine-cell techniques. We compare these results with those of published data for 3 additional stars (ϒ Cephei, β Geminorum, and β Ophiuchi), and discuss possible implications for the underlying physical mechanism(s).

High dispersion spectra for 44 southern evolved stars of spectral type K or M have been obtained. From these observations relative radial velocities of 50 m/s precision have been obtained by the method of digital cross-correlation. This method of achieving precise relative radial velocities for late-type stars, and the problems encountered in its use, are discussed. Using this method, statistically significant radial-velocity variations have been found. Two of the observed stars have their precise radial velocities presented and the potential mechanisms of their variation examined.

In view of the high incidence of duplicity among Cepheids, a large number of binary systems remain to be discovered because of the low amplitude of the orbital effect superimposed on the radial velocity variation due to radial pulsation. This fact is shown by simple statistics. The importance of combining astrometric data with radial velocity measurements is also emphasized. Finally the systematics of the radial velocity amplitude as a function of the pulsation period is considered for classical Cepheids.

Until recently the problem of collecting high resolution spectra of flare stars has been intractable since the techniques available have not been sensitive enough to reach these extremely faint objects. Although many of the nearest stars (and practically all of the nearby variable stars) belong to this class, even the ones nearest to our sun are fainter than magnitude 8 or 10. In determining the radial velocity of nearby flare stars astronomers accepted the available accuracy of ~ 1 km/s. This may be adequate for the classification of the objects into age classes (according their kinematic properties).

The other considerable group of flare stars is taken traditionally as a natural by-product of star formation processes which go on in clusters and associations. Until recently there has not been any serious attack against the widely popular hypothesis that all but a few of the flare stars discovered in the fields of stellar aggregates (their number exceeds that of the solar neighborhood flare stars) are physical members of the systems. The discovery (Szécsényi-Nagy et al. 1997, 1998) that hundreds of flare stars found in the field of M45 may not be cluster members may change the situation. Most flare stars observed there are very faint and consequently they were missing from previously published lists of Pleiades members. For one third of the objects only reliable membership probabilities have been determined, and many of them are listed as probable non-members (Haro, Chavira, & Gonzalez 1982). However, a recently published photographic proper motion survey of the Pleiades’ field (Souchay & Schilbach 1995) provided reliable membership probability values for many stars of extremely low luminosity too. Based on that about 85% of the well-documented flare stars can be – and have been – identified. Our results (Szécsényi-Nagy et al. 1997) undoubtedly prove that a substantial fraction (~ 40%) of the so called Pleiades flare stars are (more or less) definitely non-members. Since all of these new cluster membership probability calculations have been based on stellar proper motion values, in order to be able to reach a final decision, we badly need some other independent data set for the very same stars. It is to be shown that precise stellar radial velocities, an unexploited – because almost unknown – parameter for flare stars, could solve the problem by supporting or disproving these faint objects’ cluster membership. Consequently the flare stars of these two kinds (which are accidentally mixed on the photographic plates) could be classified into different age groups and their evolutionary stages and tracks could be investigated more deeply.

Our intention is to persuade astronomers involved in stellar radial velocity business that developing and using a method of high precision stellar radial-velocity measurement for late dK/dM stars is not a waste of time but a really feasible job and that we can and will contribute to the success of it by identifying the best tartgets, taking part in the necessary observations and evaluating the data.

We study the velocities of specific stellar absorption lines of Cepheid-type variable stars determining line bisector velocities from newly obtained high-resolution echelle spectra. We compare these velocities with recent CORAVEL measurements of Cepheids to reveal systematic differences of 2 – 3 kms−1.

Low-amplitude, radial-velocity variations of bright late-type stars were monitored at the 1.2-m telescope of the Dominion Astrophysical Observatory (DAO) with the hydrogen fluoride (HF) absorption-cell technique. Many of the stars appear to be semi-regular red variables (SRb and SRc) and irregular red variables (Lb). The radial-velocity amplitudes range from a few tens of meters per second to a few kilometers per second while the timescales of the variations appear to range from a few tens of days to a few hundreds of days. These irregular-looking velocities are analysed for multiperiodicities. There are also variations in the chromospheric Ca II 8662 index for a few of the variables.

I review recent progress in the field of stellar surface imaging, with particular reference to advanced methods for mapping surface-brightness distributions on magnetically active late-type stars. New signal enhancement techniques, utilising profile information from hundreds or thousands of photospheric lines simultaneously, allow images to be derived for stars several magnitudes fainter than was previously possible. For brighter stars, the same techniques make it possible to map features as small as two or three degrees in extent on the stellar surface. While this opens up whole new areas of research, such as the ability to use starspot tracking to study surface differential rotation patterns on single and binary stars, caution must be exercised in the treatment of “nuisance” parameters such as the stellar rotation rate, surface abundances and radial velocity. At the very high S:N levels we now use, the effects of systematic errors in these parameters are easier to identify, isolate and eliminate. This leads to the possibility of making precise radial velocity variations (at the few hundred m s−2 level or better) in late-type stars even with equatorial rotation speeds as high as 100 km s−1. This is particularly topical given the recent discovery that one of our prime imaging targets, the young southern K0 dwarf AB Doradus, has an astrometric companion in a highly eccentric orbit with an inferred mass close to the H-burning limit.

This paper focuses on the limitations of establishing rotation rates (to be thought of as differential radial velocities across the surface of the star) and how rotation infringes on our ability to measure precise positions of spectral lines and hence spectroscopic radial velocities.