We present new $H$ (1.5-1.8 $\mu$m) photometric and $K_1$ (1.9-2.2 $\mu$m) spectroscopic observations of the young exoplanet HD 95086 b obtained with the Gemini Planet Imager. The $H$-band magnitude has been significantly improved relative to previous measurements, whereas the low resolution $K_1$ ($\lambda/\delta\lambda \approx 66$) spectrum is featureless within the measurement uncertainties, and presents a monotonically increasing pseudo-continuum consistent with a cloudy atmosphere. By combining these new measurements with literature $L^{\prime}$ photometry, we compare the spectral energy distribution of the planet to other young planetary-mass companions, field brown dwarfs, and to the predictions of grids of model atmospheres. HD 95086 b is over a magnitude redder in $K_1-L^{\prime}$ color than 2MASS J12073346-3932539 b and HR 8799 c and d, despite having a similar $L^{\prime}$ magnitude. Considering only the near-infrared measurements, HD 95086 b is most analogous to the brown dwarfs 2MASS J2244316+204343 and 2MASS J21481633+4003594, both of which are thought to have dusty atmospheres. Morphologically, the spectral energy distribution of HD 95086 b is best fit by low temperature ($T_{\rm eff} =$ 800-1300 K), low surface gravity spectra from models which simulate high photospheric dust content. This range of effective temperatures is consistent with field L/T transition objects, but the spectral type of HD 95086 b is poorly constrained between early L and late T due to its unusual position the color-magnitude diagram, demonstrating the difficulty in spectral typing young, low surface gravity substellar objects. As one of the reddest such objects, HD 95086 b represents an important empirical benchmark against which our current understanding of the atmospheric properties of young extrasolar planets can be tested.

We present H and K band imaging polarimetry for the PDS 66 circumstellar disk obtained during the commissioning of the Gemini Planet Imager (GPI). Polarization images reveal a clear detection of the disk in to the 0.12'' inner working angle (IWA) in H band, almost 3 times as close to the star as the previous HST observations with NICMOS and STIS (0.35'' effective IWA). The centro-symmetric polarization vectors confirm that the bright inner disk detection is due to circumstellar scattered light. A more diffuse disk extends to a bright outer ring centered at 80 AU. We discuss several physical mechanisms capable of producing the observed ring + gap structure. GPI data confirm enhanced scattering on the East side of the disk which is inferred to be nearer to us. We also detect a lateral asymmetry in the South possibly due to shadowing from material within the inner working angle. This likely corresponds to a temporally variable azimuthal asymmetry observed in HST/STIS coronagraphic imaging.

We present Gemini Planet Imager (GPI) observations of AU Microscopii, a young M dwarf with an edge-on, dusty debris disk. Integral field spectroscopy and broadband imaging polarimetry were obtained during the commissioning of GPI. In our broadband imaging polarimetry observations, we detect the disk only in total intensity and find asymmetries in the morphology of the disk between the southeast and northwest sides. The southeast side of the disk exhibits a bump at 1$''$ (10 AU projected separation) that is three times more vertically extended and three times fainter in peak surface brightness than the northwest side at similar separations. This part of the disk is also vertically offset by 69$\pm$30 mas to the northeast at 1$''$ when compared to the established disk mid-plane and consistent with prior ALMA and Hubble Space Telescope/STIS observations. We see hints that the southeast bump might be a result of detecting a horizontal sliver feature above the main disk that could be the disk backside. Alternatively when including the morphology of the northwest side, where the disk mid-plane is offset in the opposite direction $\sim$50 mas between 0$.''$4 and 1$.''$2, the asymmetries suggest a warp-like feature. Using our integral field spectroscopy data to search for planets, we are 50% complete for $\sim$4 $M_\mathrm{Jup}$ planets at 4 AU. We detect a source, resolved only along the disk plane, that could either be a candidate planetary mass companion or a compact clump in the disk.

We present $H$-band observations of $\beta$ Pic with the Gemini Planet Imager's (GPI's) polarimetry mode that reveal the debris disk between ~0.3" (~6 AU) and ~1.7" (~33 AU), while simultaneously detecting $\beta$ Pic $b$. The polarized disk image was fit with a dust density model combined with a Henyey-Greenstein scattering phase function. The best fit model indicates a disk inclined to the line of sight ($\phi=85.27{\deg}^{+0.26}_{-0.19}$) with a position angle $\theta_{PA}=30.35{\deg}^{+0.29}_{-0.28}$ (slightly offset from the main outer disk, $\theta_{PA}\approx29{\deg}$), that extends from an inner disk radius of $23.6^{+0.9}_{-0.6}$ AU to well outside GPI's field of view. In addition, we present an updated orbit for $\beta$ Pic $b$ based on new astrometric measurements taken in GPI's spectroscopic mode spanning 14 months. The planet has a semi-major axis of $a=9.2^{+1.5}_{-0.4}$AU, with an eccentricity $e\leq 0.26$. The position angle of the ascending node is $\Omega=31.75{\deg}\pm0.15$, offset from both the outer main disk and the inner disk seen in the GPI image. The orbital fit constrains the stellar mass of $\beta$ Pic to $1.60\pm0.05 M_{\odot}$. Dynamical sculpting by $\beta$ Pic $b$ cannot easily account for the following three aspects of the inferred disk properties: 1) the modeled inner radius of the disk is farther out than expected if caused by $\beta$ Pic b; 2) the mutual inclination of the inner disk and $\beta$ Pic $b$ is $4{\deg}$, when it is expected to be closer to zero; and 3) the aspect ratio of the disk ($h_0 = 0.137^{+0.005}_{-0.006}$) is larger than expected from interactions with $\beta$ Pic $b$ or self-stirring by the disk's parent bodies.

We describe the astrometric calibration of the Gemini NICI Planet-Finding Campaign. The Campaign requires a relative astrometric accuracy of $\approx$ 20 mas across multi-year timescales in order to distinguish true companions from background stars by verifying common proper motion and parallax with their parent stars. The calibration consists of a correction for instrumental optical image distortion, plus on-sky imaging of astrometric fields to determine the pixel scale and image orientation. We achieve an accuracy of $\lesssim 7$ mas between the center and edge of the 18$''$ NICI field, meeting the 20 mas requirement. Most of the Campaign data in the Gemini Science Archive are accurate to this level but we identify a number of anomalies and present methods to correct the errors.

During the first-light run of the Gemini Planet Imager (GPI) we obtained K-band spectra of exoplanets HR 8799 c and d. Analysis of the spectra indicates that planet d may be warmer than planet c. Comparisons to recent patchy cloud models and previously obtained observations over multiple wavelengths confirm that thick clouds combined with horizontal variation in the cloud cover generally reproduce the planets' spectral energy distributions. When combined with the 3 to 4 um photometric data points, the observations provide strong constraints on the atmospheric methane content for both planets. The data also provide further evidence that future modeling efforts must include cloud opacity, possibly including cloud holes, disequilibrium chemistry, and super-solar metallicity.

The Gemini Planet Imager (GPI) entered on-sky commissioning and had its first-light at the Gemini South (GS) telescope in November 2013. GPI is an extreme adaptive optics (XAO), high-contrast imager and integral-field spectrograph dedicated to the direct detection of hot exo-planets down to a Jupiter mass. The performance of the apodized pupil Lyot coronagraph depends critically upon the residual wavefront error (design goal of 60 nm RMS with 5 mas RMS tip/tilt), and therefore is most sensitive to vibration (internal or external) of Gemini's instrument suite. Excess vibration can be mitigated by a variety of methods such as passive or active dampening at the instrument or telescope structure or Kalman filtering of specific frequencies with the AO control loop. Understanding the sources, magnitudes and impact of vibration is key to mitigation. This paper gives an overview of related investigations based on instrument data (GPI AO module) as well as external data from accelerometer sensors placed at different locations on the GS telescope structure. We report the status of related mitigation efforts, and present corresponding results.

The Gemini Planet Imager (GPI) entered on-sky commissioning phase, and had its First Light at the Gemini South telescope in November 2013. Meanwhile, the fast loops for atmospheric correction of the Extreme Adaptive Optics (XAO) system have been closed on many dozen stars at different magnitudes (I=4-8), elevation angles and a variety of seeing conditions, and a stable loop performance was achieved from the beginning. Ultimate contrast performance requires a very low residual wavefront error (design goal 60 nm RMS), and optimization of the planet finding instrument on different ends has just begun to deepen and widen its dark hole region. Laboratory raw contrast benchmarks are in the order of 10^-6 or smaller. In the telescope environment and in standard operations new challenges are faced (changing gravity, temperature, vibrations) that are tackled by a variety of techniques such as Kalman filtering, open-loop models to keep alignment to within 5 mas, speckle nulling, and a calibration unit (CAL). The CAL unit was especially designed by the Jet Propulsion Laboratory to control slowly varying wavefront errors at the focal plane of the apodized Lyot coronagraph by the means of two wavefront sensors: 1) a 7x7 low order Shack-Hartmann SH wavefront sensor (LOWFS), and 2) a special Mach-Zehnder interferometer for mid-order spatial frequencies (HOWFS) - atypical in that the beam is split in the focal plane via a pinhole but recombined in the pupil plane with a beamsplitter. The original design goal aimed for sensing and correcting on a level of a few nm which is extremely challenging in a telescope environment. This paper focuses on non-common path low order wavefront correction as achieved through the CAL unit on sky. We will present the obtained results as well as explain challenges that we are facing.

Using the recently installed Gemini Planet Imager (GPI), we have taken the first H-band spectrum of the planetary companion to the nearby young star beta Pictoris. GPI is designed to image and provide low-resolution spectra of Jupiter sized, self-luminous planetary companions around young nearby stars. These observations were taken covering the H-band (1.65 microns). The spectrum has a resolving power of $\sim$ 45 and demonstrates the distinctive triangular shape of a cool substellar object with low surface gravity. Using atmospheric models, we find an effective temperature of $1650 \pm 50$ K and a surface gravity of $\log(g) = 4.0 \pm 0.25$ (cgs units). These values agree well with predictions from planetary evolution models for a gas giant with mass between 10 and 12 $M_{\rm Jup}$ and age between 10 and 20 Myrs.

The Gemini Planet Imager (GPI) has as its science instrument an infrared integral field spectrograph/polarimeter (IFS). Integral field spectrographs are scientificially powerful but require sophisticated data reduction systems. For GPI to achieve its scientific goals of exoplanet and disk characterization, IFS data must be reconstructed into high quality astrometrically and photometrically accurate datacubes in both spectral and polarization modes, via flexible software that is usable by the broad Gemini community. The data reduction pipeline developed by the GPI instrument team to meet these needs is now publicly available following GPI's commissioning. This paper, the first of a series, provides a broad overview of GPI data reduction, summarizes key steps, and presents the overall software framework and implementation. Subsequent papers describe in more detail the algorithms necessary for calibrating GPI data. The GPI data reduction pipeline is open source, available from planetimager.org, and will continue to be enhanced throughout the life of the instrument. It implements an extensive suite of task primitives that can be assembled into reduction recipes to produce calibrated datasets ready for scientific analysis. Angular, spectral, and polarimetric differential imaging are supported. Graphical tools automate the production and editing of recipes, an integrated calibration database manages reference files, and an interactive data viewer customized for high contrast imaging allows for exploration and manipulation of data.

We present on-sky polarimetric observations with the Gemini Planet Imager (GPI) obtained at straight Cassegrain focus on the Gemini South 8-m telescope. Observations of polarimetric calibrator stars, ranging from nearly unpolarized to strongly polarized, enable determination of the combined telescope and instrumental polarization. We find the conversion of Stokes $I$ to linear and circular instrumental polarization in the instrument frame to be $I \rightarrow (Q_{\rm IP}, U_{\rm IP}, P_{\rm IP}, V_{\rm IP}) = (-0.037 \pm 0.010\%, +0.4338 \pm 0.0075\%, 0.4354 \pm 0.0075\%, -6.64 \pm 0.56\%)$. Such precise measurement of instrumental polarization enables $\sim 0.1\%$ absolute accuracy in measurements of linear polarization, which together with GPI's high contrast will allow GPI to explore scattered light from circumstellar disk in unprecedented detail, conduct observations of a range of other astronomical bodies, and potentially even study polarized thermal emission from young exoplanets. Observations of unpolarized standard stars also let us quantify how well GPI's differential polarimetry mode can suppress the stellar PSF halo. We show that GPI polarimetry achieves cancellation of unpolarized starlight by factors of 100-200, reaching the photon noise limit for sensitivity to circumstellar scattered light for all but the smallest separations at which the calibration for instrumental polarization currently sets the limit.

The Gemini Planet Imager (GPI) is an instrument designed to directly image planets and circumstellar disks from 0.9 to 2.5 microns (the $YJHK$ infrared bands) using high contrast adaptive optics with a lenslet-based integral field spectrograph. We develop an extraction algorithm based on a least-squares method to disentangle the spectra and systematic noise contributions simultaneously. We utilize two approaches to adjust for the effect of flexure of the GPI optics which move the position of light incident on the detector. The first method is to iterate the extraction to achieve minimum residual and the second is to cross-correlate the detector image with a model image in iterative extraction steps to determine an offset. Thus far, this process has made clear qualitative improvements to the cube extraction by reducing the Moir\'{e} pattern. There are also improvements to the automated routines for finding flexure offsets which are reliable to with $\sim0.5$ pixel accuracy compared to pixel accuracy prior. Further testing and optimization will follow before implementation into the GPI pipeline.

The Gemini Planet Imager (GPI) Extreme Adaptive Optics Coronograph contains an interferometric mode: a 10-hole non-redundant mask (NRM) in its pupil wheel. GPI operates at $Y, J, H$, and $K$ bands, using an integral field unit spectrograph (IFS) to obtain spectral data at every image pixel. NRM on GPI is capable of imaging with a half resolution element inner working angle at moderate contrast, probing the region behind the coronagraphic spot. The fine features of the NRM PSF can provide a reliable check on the plate scale, while also acting as an attenuator for spectral standard calibrators that would otherwise saturate the full pupil. NRM commissioning data provides details about wavefront error in the optics as well as operations of adaptive optics control without pointing control from the calibration system. We compare lab and on-sky results to evaluate systematic instrument properties and examine the stability data in consecutive exposures. We discuss early on-sky performance, comparing images from integration and tests with the first on-sky images, and demonstrate resolving a known binary. We discuss the status of NRM and implications for future science with this mode.

The Gemini Planet Imager is an extreme AO instrument with an integral field spectrograph (IFS) operating in Y, J, H, and K bands. Both the Gemini telescope and the GPI instrument are very complex systems. Our goal is that the combined telescope and instrument system may be run by one observer operating the instrument, and one operator controlling the telescope and the acquisition of light to the instrument. This requires a smooth integration between the two systems and easily operated control interfaces. We discuss the definition of the software and hardware interfaces, their implementation and testing, and the integration of the instrument with the telescope environment.

The Gemini Planet Imager (GPI) is a next-generation, facility instrument currently being commissioned at the Gemini South observatory. GPI combines an extreme adaptive optics system and integral field spectrograph (IFS) with an apodized-pupil Lyot coronagraph (APLC) producing an unprecedented capability for directly imaging and spectroscopically characterizing extrasolar planets. GPI's operating goal of $10^{-7}$ contrast requires very precise alignments between the various elements of the coronagraph (two pupil masks and one focal plane mask) and active control of the beam path throughout the instrument. Here, we describe the techniques used to automatically align GPI and maintain the alignment throughout the course of science observations. We discuss the particular challenges of maintaining precision alignments on a Cassegrain mounted instrument and strategies that we have developed that allow GPI to achieve high contrast even in poor seeing conditions.

The Gemini Planet Imager (GPI) is a complex optical system designed to directly detect the self-emission of young planets within two arcseconds of their host stars. After suppressing the starlight with an advanced AO system and apodized coronagraph, the dominant residual contamination in the focal plane are speckles from the atmosphere and optical surfaces. Since speckles are diffractive in nature their positions in the field are strongly wavelength dependent, while an actual companion planet will remain at fixed separation. By comparing multiple images at different wavelengths taken simultaneously, we can freeze the speckle pattern and extract the planet light adding an order of magnitude of contrast. To achieve a bandpass of 20%, sufficient to perform speckle suppression, and to observe the entire two arcsecond field of view at diffraction limited sampling, we designed and built an integral field spectrograph with extremely low wavefront error and almost no chromatic aberration. The spectrograph is fully cryogenic and operates in the wavelength range 1 to 2.4 microns with five selectable filters. A prism is used to produce a spectral resolution of 45 in the primary detection band and maintain high throughput. Based on the OSIRIS spectrograph at Keck, we selected to use a lenslet-based spectrograph to achieve an rms wavefront error of approximately 25 nm. Over 36,000 spectra are taken simultaneously and reassembled into image cubes that have roughly 192x192 spatial elements and contain between 11 and 20 spectral channels. The primary dispersion prism can be replaced with a Wollaston prism for dual polarization measurements. The spectrograph also has a pupil-viewing mode for alignment and calibration.

An Atmospheric Dispersion Corrector (ADC) uses a double-prism arrangement to nullify the vertical chromatic dispersion introduced by the atmosphere at non-zero zenith distances. The ADC installed in the Gemini Planet Imager (GPI) was first tested in August 2012 while the instrument was in the laboratory. GPI was installed at the Gemini South telescope in August 2013 and first light occurred later that year on November 11th. In this paper, we give an overview of the characterizations and performance of this ADC unit obtained in the laboratory and on sky, as well as the structure of its control software.

The Gemini Planet Imager (GPI) is a new facility, extreme adaptive optics (AO), coronagraphic instrument, currently being integrated onto the 8-meter Gemini South telescope, with the ultimate goal of directly imaging extrasolar planets. To achieve the contrast required for the desired science, it is necessary to quantify and mitigate wavefront error (WFE). A large source of potential static WFE arises from the primary AO wavefront sensor (WFS) detector's use of multiple readout segments with independent signal chains including on-chip preamplifiers and external amplifiers. Temperature changes within GPI's electronics cause drifts in readout segments' bias levels, inducing an RMS WFE of 1.1 nm and 41.9 nm over 4.44 degrees Celsius, for magnitude 4 and 11 stars, respectively. With a goal of $<$2 nm of static WFE, these are significant enough to require remedial action. Simulations imply a requirement to take fresh WFS darks every 2 degrees Celsius of temperature change, for a magnitude 6 star; similarly, for a magnitude 7 star, every 1 degree Celsius of temperature change. For sufficiently dim stars, bias drifts exceed the signal, causing a large initial WFE, and the former periodic requirement practically becomes an instantaneous/continuous one, making the goal of $<$2 nm of static WFE very difficult for stars of magnitude 9 or fainter. In extreme cases, this can cause the AO loops to destabilize due to perceived nonphysical wavefronts, as some of the WFS's Shack-Hartmann quadcells are split between multiple readout segments. Presented here is GPI's AO WFS geometry, along with detailed steps in the simulation used to quantify bias drift related WFE, followed by laboratory and on sky results, and concluded with possible methods of remediation.

We present the first results from the polarimetry mode of the Gemini Planet Imager (GPI), which uses a new integral field polarimetry architecture to provide high contrast linear polarimetry with minimal systematic biases between the orthogonal polarizations. We describe the design, data reduction methods, and performance of polarimetry with GPI. Point spread function subtraction via differential polarimetry suppresses unpolarized starlight by a factor of over 100, and provides sensitivity to circumstellar dust reaching the photon noise limit for these observations. In the case of the circumstellar disk around HR 4796A, GPI's advanced adaptive optics system reveals the disk clearly even prior to PSF subtraction. In polarized light, the disk is seen all the way in to its semi-minor axis for the first time. The disk exhibits surprisingly strong asymmetry in polarized intensity, with the west side >9 times brighter than the east side despite the fact that the east side is slightly brighter in total intensity. Based on a synthesis of the total and polarized intensities, we now believe that the west side is closer to us, contrary to most prior interpretations. Forward scattering by relatively large silicate dust particles leads to the strong polarized intensity on the west side, and the ring must be slightly optically thick in order to explain the lower brightness in total intensity there. These findings suggest that the ring is geometrically narrow and dynamically cold, perhaps shepherded by larger bodies in the same manner as Saturn's F ring.

The Gemini Planet Imager instrument's adaptive optics (AO) subsystem was designed specifically to facilitate high-contrast imaging. It features several new technologies, including computationally efficient wavefront reconstruction with the Fourier transform, modal gain optimization every 8 seconds, and the spatially filtered wavefront sensor. It also uses a Linear-Quadratic-Gaussian (LQG) controller (aka Kalman filter) for both pointing and focus. We present on-sky performance results from verification and commissioning runs from December 2013 through May 2014. The efficient reconstruction and modal gain optimization are working as designed. The LQG controllers effectively notch out vibrations. The spatial filter can remove aliases, but we typically use it oversized by about 60% due to stability problems.

We present the high-contrast image processing techniques used by the Gemini NICI Planet-Finding Campaign to detect faint companions to bright stars. NICI (Near Infrared Coronagraphic Imager) is an adaptive optics instrument installed on the 8-m Gemini South telescope, capable of angular and spectral difference imaging and specifically designed to image exoplanets. The Campaign data pipeline achieves median contrasts of 12.6 magnitudes at 0.5" and 14.4 magnitudes at 1" separation, for a sample of 45 stars (V= 4.3-13.9 mag) from the early phase of the Campaign. We also present a novel approach to calculating contrast curves for companion detection based on 95% completeness in the recovery of artificial companions injected into the raw data, while accounting for the false-positive rate. We use this technique to select the image processing algorithms that are more successful at recovering faint simulated point sources. We compare our pipeline to the performance of the LOCI algorithm for NICI data and do not find significant improvement with LOCI.

We report results of a direct imaging survey for giant planets around 80 members of the Beta Pic, TW Hya, Tucana-Horologium, AB Dor, and Hercules-Lyra moving groups, observed as part of the Gemini NICI Planet-Finding Campaign. For this sample, we obtained median contrasts of \Delta H=13.9 mag at 1" in combined CH4 narrowband ADI+SDI mode and median contrasts of \Delta H=15.1 mag at 2" in H-band ADI mode. We found numerous (>70) candidate companions in our survey images. Some of these candidates were rejected as common-proper motion companions using archival data; we reobserved with NICI all other candidates that lay within 400 AU of the star and were not in dense stellar fields. The vast majority of candidate companions were confirmed as background objects from archival observations and/or dedicated NICI campaign followup. Four co-moving companions of brown dwarf or stellar mass were discovered in this moving group sample: PZ Tel B (36+-6 MJup, 16.4+-1.0 AU, Biller et al. 2010), CD -35 2722B (31+-8 MJup, 67+-4 AU, Wahhaj et al. 2011), HD 12894B (0.46+-0.08 MSun, 15.7+-1.0 AU), and BD+07 1919C (0.20+-0.03 MSun, 12.5+-1.4 AU). From a Bayesian analysis of the achieved H band ADI and ASDI contrasts, using power-law models of planet distributions and hot-start evolutionary models, we restrict the frequency of 1--20 MJup companions at semi-major axes from 10--150 AU to <18% at a 95.4% confidence level using DUSTY models and to <6% at a 95.4% using COND models.

We have carried out high contrast imaging of 70 young, nearby B and A stars to search for brown dwarf and planetary companions as part of the Gemini NICI Planet-Finding Campaign. Our survey represents the largest, deepest survey for planets around high-mass stars (~1.5-2.5 M_sun) conducted to date and includes the planet hosts beta Pic and Fomalhaut. We obtained follow-up astrometry of all candidate companions within 400 AU projected separation for stars in uncrowded fields and identified new low-mass companions to HD 1160 and HIP 79797. We have found that the previously known young brown dwarf companion to HIP 79797 is itself a tight (3 AU) binary, composed of brown dwarfs with masses 58 (+21, -20) M_Jup and 55 (+20, -19) M_Jup, making this system one of the rare substellar binaries in orbit around a star. Considering the contrast limits of our NICI data and the fact that we did not detect any planets, we use high-fidelity Monte Carlo simulations to show that fewer than 20% of 2 M_sun stars can have giant planets greater than 4 M_Jup between 59 and 460 AU at 95% confidence, and fewer than 10% of these stars can have a planet more massive than 10 M_Jup between 38 and 650 AU. Overall, we find that large-separation giant planets are not common around B and A stars: fewer than 10% of B and A stars can have an analog to the HR 8799 b (7 M_Jup, 68 AU) planet at 95% confidence. We also describe a new Bayesian technique for determining the ages of field B and A stars from photometry and theoretical isochrones. Our method produces more plausible ages for high-mass stars than previous age-dating techniques, which tend to underestimate stellar ages and their uncertainties.

Hipparcos astrometric binaries were observed with the NICI adaptive optics system at Gemini-S, completing the work of Paper I (Tokovinin et al. 2012). Among the 65 F, G, K dwarfs within 67pc of the Sun studied here, we resolve 18 new sub-arcsecond companions, re-measure 7 known astrometric pairs, and establish the physical nature of yet another three wider companions. The 107 astrometric binaries targeted at Gemini so far have 38 resolved companions with separations under 3". Modeling shows that bright enough companions with separations on the order of an arcsecond can perturb the Hipparcos astrometry when they are not accounted for in the data reduction. However, the resulting bias of parallax and proper motion is generally below formal errors and such companions cannot produce fake acceleration. This work contributes to the multiplicity statistics of nearby dwarfs by bridging the gap between spectroscopic and visual binaries and by providing estimates of periods and mass ratios for many astrometric binaries.

A subset of 51 Hipparcos astrometric binaries among FG dwarfs within 67pc has been surveyed with the NICI adaptive optics system at Gemini-S, directly resolving for the first time 17 sub-arcsecond companions and 7 wider ones. Using these data together with published speckle interferometry of 57 stars, we compare the statistics of resolved astrometric companions with those of a simulated binary population. The fraction of resolved companions is slightly lower than expected from binary statistics. About 10% of astrometric companions could be "dark" (white dwarfs and close pairs of late M-dwarfs). To our surprise, several binaries are found with companions too wide to explain the acceleration. Re-analysis of selected intermediate astrometric data shows that some acceleration solutions in the original Hipparcos catalog are spurious.