MID-INFRARED HIGH-CONTRAST IMAGING OF HD 114174 B: AN APPARENT AGE DISCREPANCY IN A "SIRIUS-LIKE" BINARY SYSTEM

Spatially resolved binaries provide a unique opportunity to study the composition, age, and evolution of compact objects by comparing their properties to those of the primary star. In particular, white dwarf (WD) companions orbiting nearby stars, such as Sirius and Procyon, lend themselves to detailed analysis, particularly when the metallicity and evolutionary state of the primary can be determined. Through gravitational interactions, binaries also permit calculation of a dynamical mass (Boden et al. 2006). Model-independent masses may then be used to test and calibrate theoretical WD cooling models, atmospheric models, and calculations of initial-to-final mass ratios (Isern et al. 2007). Directly imaged WD companions orbiting main-sequence stars (spectral type K or earlier) are, however, scarce, constituting ∼8% of the local sample of 132 WDs discovered within 20 pc of the Sun (Holberg 2009). Presently only five "Sirius-like systems" (Sirius, Procyon, 40 Eridani, HR 5692, and 56 Perseus) have fully characterized orbits (Holberg et al. 2013).

A new Sirius-like system, HD 114174, was recently discovered as part of the TRENDS high contrast imaging program (Crepp et al. 2012b, 2013). The primary star, HD 114174 A (Table 1), shows a long-term radial velocity (RV) acceleration of ∼61 m s⁻¹ yr⁻¹ over a 16 yr time baseline. The companion's flux and neutral color, J − K = 0.12  ±  0.16, originally led Crepp et al. (2013) to classify the object as an ≈T3 brown dwarf. However, a lower limit of 0.26 ± 0.01 M_☉ on the companion's mass derived from a combination of RV data and imaging led them to re-evaluate HD 114174 B as a compact object.

Table 1. Properties of HD 114174 A from Crepp et al. (2013)

Right ascension (J2000)	13 08 51.02
Declination (J2000)	+05 12 26.06
B	7.47 ± 0.01
V	6.8 ± 0.001
R	6.3 ± 0.02
I	6.0 ± 0.01
Ya	5.71 ± 0.01
J	5.613 ± 0.026
H	5.312 ± 0.027
Ks	5.202 ± 0.023
Distance (pc)	26.14 ± 0.37
Mass (M☉)	1.05 ± 0.05
Radius (R☉)	1.06
Luminosity (L☉)	1.13
Age (isochonal) (Gyr)	4.7$^{+2.3}_{-2.6}$
Age (gyrochronological) (Gyr)	4.0$^{+0.96}_{-1.09}$
(Fe/H)	0.07 ± 0.03
log(g) (cm s−2)	4.51 ± 0.06
Teff (K)	5781 ± 44
Spectral type	G5 IV-V
v sin(i) (km/s)	1.8 ± 0.5

Note. ^aWe estimate the Y-band magnitude of HD 114174 A by fitting a blackbody function to BVRIJHK_s photometric measurements (Section 3.2).

Download table as:  ASCIITypeset image

Initial characterization of HD 114174 B relied on Keck adaptive optics (AO) observations (Wizinowich et al. 2000) taken in the J and K bands using NIRC2 (instrument PI: Keith Matthews). Crepp et al. (2013) attempted to recover the companion in the L'-band at Keck, but were only able to place an upper limit on its brightness based on a non-detection. Their photometric observations were used along with WD atmospheric models to characterize the newly found companion. Since only two data points and one upper limit were available, the resulting effective temperature (T_eff = 8100  ±  4000 K) was poorly constrained, leaving the physical parameters of the WD uncertain.

We have acquired new photometric observations of HD 114174 B in the L'-band (λ_c = 3.7 μm) using the Large Binocular Telescope (LBT). The LBT AO system delivers high-fidelity wavefront correction through the use of a pyramid wavefront sensor, which offers improved noise characteristics compared to a Shack–Hartmann design, and a deformable secondary mirror (672 actuators) that provides fine spatial correction while also lowering thermal background levels at mid-infrared wavelengths (Esposito et al. 2011; Tozzi et al. 2008; Close et al. 2012). These design features enable the LBT AO system to achieve unprecedented contrast levels (Skemer et al. 2012).

In this Letter, we present the first imaging recovery of HD 114174 B in the L'-band (Figure 1). Our new observations obtained with the LBT Interferometer (LBTI; Hinz et al. 2008) and L/M-band InfraRed Camera (LMIRCam; Skrutskie et al. 2010) rank amongst the deepest high contrast imaging in this band to date, and represent the deepest inside of 1'' (ΔL' = 10.15  ±  0.15 ρ = 0675 ± 0016). HR 8799 b (ΔL' = 10.42  ±  0.2), GJ 758 B (ΔL' = 11.3  ±  0.14), and GJ 504 b (ΔL' = 12.9  ±  0.19) have been detected with greater contrast, but reside at larger separations from their primary stars: 172, 182, and 249, respectively (Marois et al. 2008; Janson et al. 2011; Kuzuhara et al. 2013). Additional photometric measurements⁶ establish a more complete spectral energy distribution (SED) thus reducing uncertainty in HD 114174 B's physical properties (Table 2). In particular, our new L'-band data lowers the effective temperature by ∼3800 K and reduces the relative uncertainty from 49% to 8%. A cooler effective temperature corresponds to an older WD. Assuming that HD 114174 A,B have co-evolved, the new cooling age of the WD companion reveals an apparent discrepancy in their ages (2.3σ). We also provide a new astrometric data point that will assist with on-going efforts to constrain the orbit and dynamical mass of the WD companion.

Zoom In Zoom Out Reset image size

Table 2. Photometry and Best Fit Hydrogen Model for HD 114174 B

ΔY	>8.2 ± 0.7 mag
ΔJ	10.48 ± 0.11 mag
ΔK	10.75 ± 0.12 mag
ΔL'	10.15 ± 0.15 mag
Y	>14.0 ± 0.7
J	16.06 ± 0.11
Ks	15.94 ± 0.12
L'	15.30 ± 0.16
MY	>11.85 ± 0.7
MJ	13.98 ± 0.11
$M_{K_{s}}$	13.85 ± 0.12
ML	13.21 ± 0.16
Mass (M☉)	0.54 ± 0.01
Radius (R☉)	0.01301 ± 0.00019
log L/L☉	−4.30 ± 0.02
Age (Gyr)	7.77 ± 0.24
log g	7.94 ± 0.03
Teff (K)	4260 ± 360
Mv	7.90 ± 0.03
Mbol	15.49 ± 0.03

Download table as:  ASCIITypeset image

We observed HD 114174 on UT 2013 May 24 with LBTI (Hinz et al. 2008) and LMIRCam (Skrutskie et al. 2010; Leisenring et al. 2012). While LBTI is capable of imaging with both telescope dishes simultaneously, we only used the left side due to a cooling incident that temporarily disabled the right-side adaptive secondary mirror. Conditions were photometric with ∼09 seeing. We obtained 4500 images with an integration time of 0.291 s ×3 coadds each at λ = 3.8 μm (L'). Shortly before the observation, the LMIRCam electronics suffered from an overheating incident (now corrected) that corrupted alternating frames. As a result, our final data set comprises ∼30 minutes of on-sky data taken over an 80 minute time-frame. We did not use a coronagraph to occult the primary star.

LBTI sits at a bent Gregorian focus with no rotator, so all data is taken in "pupil-tracking mode" by default, where astronomical fields rotate with changing parallactic angle. For high-contrast imaging, this allows data to be processed by Angular Differential Imaging (ADI; Marois et al. 2006) and derivative algorithms such as Locally Optimized Combination of Images (Lafrenière et al. 2007; Pueyo et al. 2012) and principal component analysis (PCA) (Soummer et al. 2012; Amara & Quanz 2012). Over the course of our observations, the parallactic angle changed by 33°.

For each frame, we interpolated over bad pixels, nod-subtracted, removed residual column bias noise, and binned the data (2 × 2) to further suppress bad pixels. We then registered the images and selected the best 80% of frames for use by cross-correlation. Finally, we coadded the images by combining 20 frames to reduce the necessary processing for our high-contrast imaging algorithm.

We used PCA to subtract an average point-spread function (PSF) from each image, and then modeled the variability of the residuals to remove the largest variable components. The residual images were then rotated and combined following the basic steps of ADI. Our combined image is shown in Figure 1. HD 114174 B is detected with a signal-to-noise of 10.

3.1. L'-band Photometry

3.2. Y-band Upper-limit from Non-detection

3.3. L'-band Astrometry

To calibrate the L' LBT astrometry measurements, we observed the Washington Double Star (WDS) Catalog⁷ standard, 70 Oph, which is classified with the WDS high grade for ephemerides accuracy (Hartkopf et al. 2001; Eggenberger et al. 2008). From three different observations dithering around the chip, we find that LMIRCam's plate-scale is 10.68  ±  0.02 mas pixel⁻¹ and true north is counterclockwise from vertical by 0309  ±  0199. Applying these results to HD 114174 B, we find that the B component is separated from HD 114174 A by 0675  ±  0016 with a position angle of 1743  ±  14.

Figure 2 shows the astrometry for HD 114174 B taken over a 2.25 yr baseline. The first four points were taken with Keck/NIRC2 while the last is from LBT/LMIRCam. The combined observations self-consistently show systemic orbital motion, although the astrometry does not yet exhibit sufficient curvature to determine orbital parameters (e.g., Crepp et al. 2012a). Continued Doppler and astrometric observations are needed to calculate a model independent mass for this benchmark WD.

Zoom In Zoom Out Reset image size

4.1. WD Atmospheric Composition

Details of our model fitting procedure are described in Crepp et al. (2013) and Bergeron et al. (2001). We compare models with a pure hydrogen atmosphere, a pure helium atmosphere, as well as a range of mixed atmospheres. We used the theoretical models developed by Tremblay et al. (2011) for the hydrogen atmosphere, and those of Bergeron et al. (2011) for the pure helium model atmosphere. Mixed atmospheric models are based on the calculations of Bergeron & Leggett (2002).

Figure 3 shows the best model fits for our photometric data using pure atmospheres. While Crepp et al. (2013) were unable to ascertain HD 114174 B's composition, our new observations and model fits suggest that HD 114174 B has a hydrogen-rich atmosphere. Both the pure hydrogen and pure helium models produce reasonable fits to the observed photometry. However, we obtain χ² = 2.74 for the hydrogen solution and χ² = 6.03 for the helium solution. A key discriminator between the models is a precise parallax measurement from Hipparcos, π = 38.27 ± 0.24 mas, which allows for an accurate radius calculation for each effective temperature.

Zoom In Zoom Out Reset image size

Accounting for the L'-band photometric observation, which greatly expands wavelength coverage, we find T_eff = 4260 ± 360 K, i.e., lower than the (T_eff = 8100 ± 4000 K) range reported in Crepp et al. (2013). Our new Y-band upper limit does not place any significant constraint on the models but is consistent with the solution. Table 2 displays the resulting physical properties obtained from our new hydrogen model fit for the WD companion. Using the initial-to-final mass relations from Williams et al. (2009) we find that HD 114174 B had an initial mass M_prog = 1.58  ±  0.17 M_☉ with the hydrogen solution, and M_prog = 3.98  ±  0.16M_☉ with the helium variant.

We also consider mixed hydrogen/helium atmosphere, which have been shown to reproduce the SEDs of many cool WDs (e.g., Bergeron et al. 2001; Kilic et al. 2010), see however Kowalski & Saumon 2006. We performed fits using a grid of mixed model atmospheres that cover a range in He abundance from log(He/H) = −2 to log(He/H) = 8. Qualitatively, the results of these fits are as good as those from pure H and He models. Examination of how χ² varies as a function of log(He/H) reveals several local minima, demonstrating that there is simply too much degeneracy in the solutions from the mixed model grid to constrain the composition. Similar to the pure H and He fits, the optical/UV flux varies considerably as a function of log(He/H), reinforcing the need for optical and/or UV data to determine HD 114174 B's composition.

While it is possible to acquire additional measurements in the M-band (λ_c = 4.78 μm) and possibly R-band (λ_c = 0.65 μm) using speckle-imaging (Howell et al. 2011), our SED fits indicate that these wavelengths would provide little leverage to further understand HD 114174 B (see Figure 3). Note that the flux at λ = 0.4 μm differs significantly between the hydrogen and helium models; the hydrogen models predict a flux of 0.045 mJy whereas the helium models predict a flux of 0.209 mJy. High-resolution space-based optical/UV imaging and spectroscopy would thus permit a definitive determination of HD 114174 B's atmospheric composition.

4.2. An Apparent Age Discrepancy

Recent advances in AO correction and wavefront sensor technology (Esposito et al. 2011) combined with infrared-optimized instrumentation available at the LBT (Hinz et al. 2008; Skrutskie et al. 2010) provide a unique platform for high-contrast imaging studies in the 3–5 μm wavelength range. We have acquired new L' observations of the faint WD companion orbiting HD 114174, a purported benchmark WD discovered by the TRENDS high-contrast imaging survey (Crepp et al. 2013). HD 114174 B was imaged previously at near-infrared wavelengths, but observations from Keck using NIRC2 resulted in a non-detection in the L'-band. Using LMIRCam at the LBT, we recover the companion at 10σ. HD 114174 B has an apparent magnitude L' = 15.3  ±  0.16 mag and is ΔL' = 10.15  ±  0.15 mag fainter than its parent star. Our observations constitute the largest ΔL' high-contrast imaging detection at subarcsecond separations to date.

By extending the wavelength coverage for this "Sirius-like" WD, we are able to place stronger constraints on its physical properties compared to the discovery paper. Our observations support the interpretation that HD 114174 B is a compact object with neutral infrared colors (J − L' = 0.76  ±  0.19 mag, K − L' = 0.64  ±  0.20). New model fits suggest that a pure hydrogen atmosphere is preferred over a pure helium atmosphere, although mixed atmospheres provide similar results. Irrespective of its chemical composition, we confirm that HD 114174 B is an old and cool benchmark WD.

For plausible mixtures of hydrogen and helium, we find that the effective temperature of HD 114174 B is significantly lower than originally determined by Crepp et al. (2013) and hence our updated cooling age is longer. However, a longer cooling age (t_c = 7.77  ±  0.24 Gyr) is inconsistent with the age of HD 114174 A (t = 4.7$^{+2.3}_{-2.6}$ Gyr from isochrones, t = 4.0$^{+0.96}_{-1.09}$ Gyr from gyrochronology). This disagreement is exacerbated by the lower estimated progenitor mass compared with Crepp et al. (2013): a M_prog = 1.58  ±  0.17 M_☉ star will have a substantial (∼3.2 Gyr) main sequence lifetime. This apparent discrepancy could be the result of limited photometric data combined with inadequacies in WD models at near-IR wavelengths, or it may be the manifestation of a physical process involving winds from the WD progenitor that increase the spin rate of the companion star. Diffraction-limited imaging and spectroscopic measurements at optical and UV wavelengths from space using Hubble Space Telescope/STIS would provide vital leverage for characterizing HD 114174 B's atmosphere (see Figure 3). Such observations would eliminate the ambiguity in the WD's effective temperature, which in turn would refine HD 114174 B's cooling age, shedding light on the history and properties of this benchmark system.

C.T.M. and J.R.C. acknowledge support from NASA origins of solar systems grant NNX13AB03G. M.K. acknowledges the support of the NSF under grant AST-1312678. A.S. is supported by NASA Origins of Solar Systems grant NNX13AJ17G. The LBT is an international collaboration among institutions in the United States, Italy and Germany. LBT Corporation partners are: The University of Arizona on behalf of the Arizona university system; Istituto Nazionale di Astrofisica, Italy; LBT Beteiligungsgesellschaft, Germany, representing the Max-Planck Society, the Astrophysical Institute Potsdam, and Heidelberg University; The Ohio State University, and The Research Corporation, on behalf of The University of Notre Dame, University of Minnesota and University of Virginia. The LBT Interferometer is funded by the NASA as part of its Exoplanet Exploration program. LMIRCam is funded by the NSF through grant AST-0705296.