Dynamical Masses for the Hyades Binary System vB 120

We report spectroscopic observations of vB 120 (HD 30712), a 5.7 yr astrometric–spectroscopic binary system in the Hyades cluster. We combine our radial velocities with others from the literature, and with existing speckle interferometry measurements, to derive an improved 3D orbit for the system. We infer component masses of M 1 = 1.065 ± 0.018 M ☉ and M 2 = 1.008 ± 0.016 M ☉, and an orbital parallax of 21.86 ± 0.15 mas, which we show to be more accurate than the parallax from Gaia DR3. This is the ninth binary or multiple system in the Hyades with dynamical mass determinations, and one of the examples with the highest precision. An analysis of its spectral energy distribution yields the absolute radii of the stars, R 1 = 0.968 ± 0.012 R ☉ and R 2 = 0.878 ± 0.013 R ☉, and effective temperatures of 5656 ± 56 K and 5489 ± 60 K for the primary and secondary, respectively. A comparison of these properties with the predictions of current stellar evolution models for the known age and metallicity of the cluster shows only minor differences.


INTRODUCTION
Detached binary systems in which the dynamical masses of the components can be determined accurately and precisely have long been used to provide stringent tests of stellar evolution theory.For binaries belonging to a cluster of known age and chemical composition, the constraint is much stronger because one is no longer permitted to adjust those properties freely (except within their uncertainties) in order to reach the best agreement between the models and the observations.The Hyades is among the clusters with the most binary systems for which masses have been measured (eight to date).One of those systems (vB 22 = HD 27130;Brogaard et al. 2021, and references therein) is eclipsing, and therefore even more useful because it enables the absolute radii to be determined. 1Another relies only on astrometry (vB 80 = HD 28485; Torres 2019).The others are all astrometricspectroscopic systems.These have advantages of their own, as they provide a model-independent estimate of the distance via the orbital parallax.
Here we report an orbital analysis for another Hyades binary, vB 120 (HD 30712, V = 7.73, G5 V), a recognized member of the cluster from its proper motion, radial velocity (RV), and parallax.The binary nature of vB 120 was announced by Griffin et al. (1988), based on spectroscopic observations started in 1973.The pair was first resolved spatially by the technique of speckle interferometry in 1985interferometry in (McAlister et al. 1987)), and subsequently also by others.Griffin (2012) presented the first doublelined spectroscopic orbit for the system, featuring a small 1 One other double-lined eclipsing binary in the Hyades with dynamical mass determinations, V471 Tau, is a post-common envelope system consisting of a K-type main-sequence star and a DA white dwarf (e.g., Muirhead et al. 2022).Because of the prior interaction of the components, it is not suitable for testing models, as it is not representative of single-star evolution.
eccentricity and a period near 5.7 yr.An astrometric orbit that also used the RVs from Griffin was published by Tokovinin et al. (2015), from which the total mass of the binary was reported as 2.1 M ⊙ .Docobo et al. (2018) incorporated a few more speckle measurements, and inferred the individual masses by adopting the spectroscopic elements of Griffin.The mass uncertainties were fairly large (∼15%), however, and were limited mostly by the astrometric orbit, which suffered from incomplete phase coverage and a few obvious outliers.
In the interim, several more speckle measurements have become available, and additionally vB 120 was monitored at the Center for Astrophysics (CfA) for 14 yr, as part of a long-running spectroscopic survey of several hundred stars in the Hyades region.This presents an opportunity to significantly improve the mass determinations, through an updated astrometric-spectroscopic orbital analysis.We also aim to infer the component radii and effective temperatures, from an analysis of the spectral energy distribution aided by the orbital parallax.
The paper is structured as follows.Our RV measurements of vB 120 are presented in Section 2. The astrometric observations from the literature are described in Section 3, and Section 4 gives the details of our joint spectroscopic-astrometric orbital analysis.The fit to the spectral energy distribution of vB 120 is shown in Section 5.The photometric variability of the system is discussed in Section 6, along with other measures of stellar activity.The properties we derive for the system (masses, absolute magnitudes, temperatures, radii) are compared against current stellar evolution models in Section 7, and our conclusions may be found in Section 8.

SPECTROSCOPIC OBSERVATIONS
vB 120 was observed spectroscopically at the CfA between 1992 October and 2003 February, with the Digital Speedometer (Latham 1992) on the 1.5m Wyeth reflec-tor at the (now closed) Oak Ridge Observatory (Massachusetts, USA).This instrument delivered a resolving power of R ≈ 35, 000, and was equipped with a photoncounting Reticon detector that recorded a single echelle order 45 Å wide centered near 5187 Å.The main spectral feature in this region is the Mg I b triplet.Signalto-noise ratios for the 40 usable exposures range between 11 and 29 per resolution element of 8.5 km s −1 .Wavelength solutions relied on exposures of a thorium-argon lamp taken before and after each science exposure.The zeropoint of the instrument was monitored by means of exposures of the twilight sky in the evening and morning.Those observations were used to calculate and apply small (typically ≤ 2 km s −1 ) run-to-run corrections to the raw velocities we describe next, in order to place them on a uniform system (see Latham 1992).This native CfA system is slightly offset from the IAU reference frame by 0.14 km s −1 (Stefanik et al. 1999), as determined from observations of minor planets in the solar system.In order to remove this shift, we adjusted the velocities by adding +0.14 km s −1 .
Radial velocities for both components were measured with the two-dimensional cross-correlation algorithm TODCOR (Zucker & Mazeh 1994).We used synthetic templates from a large library of pre-computed spectra based on model atmospheres by R. L. Kurucz, and a line list manually adjusted to provide a better match to real stars (see Nordström et al. 1994;Latham et al. 2002).These models adopt a microturbulent velocity of ξ = 2 km s −1 , along with a macroturbulent velocity of ζ RT = 1 km s −1 .Grids of correlations for a range of template parameters gave the highest average correlation for effective temperatures of 5500 K and a total line broadening of 6 km s −1 for both stars.The line broadening is dominated by rotation (v sin i), but includes any difference in the true macroturbulent velocity compared to the value built into the templates.Surface gravities were held at log g = 4.5 for both stars, and the metallicity was solar, which is close to the composition of the Hyades ([Fe/H] = +0.18;Dutra-Ferreira et al. 2016, see also Section 7).
In our experience, the narrow spectral window of the observations can occasionally produce systematic errors in the velocities, which are caused by lines of the components shifting in and out of the spectral order in opposite directions as a function of orbital phase.We corrected these effects through numerical simulations (see Latham et al. 1996;Torres et al. 1997).The corrections were smaller than 0.2 km s −1 , on average, but were applied nonetheless.Table 1 lists the final velocities and their formal uncertainties.
More precise temperatures for the vB 120 components were obtained by interpolation among the templates, following Torres et al. (2002).This resulted in values of 5620 and 5450 K for the primary and secondary, with estimated uncertainties of 100 K.As the true metallicity of the cluster is slightly supersolar, we repeated this exercise for [Fe/H] = +0.5, which gave temperatures about 300 K hotter for each star.Interpolation to [Fe/H] = +0.18then led to final values of 5730 and 5560 K (±100 K).
Using TODCOR, we obtained a flux ratio between the components of ℓ 2 /ℓ 1 = 0.664 ± 0.026.This corresponds to a magnitude difference of ∆m = 0.44 ± 0.04 at the wavelength of our observations.We use this estimate later in Section 5.
In addition to our own RVs, the orbital analysis described below incorporated the 41 pairs of velocities of Griffin (2012), which are of similar precision as ours.We adopted relative weights for those measurements as recommended by the author, along with the specified error for an observation of unit weight.Separate spectroscopic orbital solutions using the Griffin observations and our own give similar velocity semiamplitudes.

ASTROMETRIC MEASUREMENTS
Speckle observations of vB 120 have been recorded by several authors since it was first resolved in 1985.A listing of all measurements from the Washington Double Star Catalog (WDS; Worley & Douglass 1997;Mason et al. 2001) was kindly provided by R. Matson (U.S. Naval Observatory).The double star designation in this catalog is WDS J04506+1505AB.While many of the early measurements suffer from a 180 • ambiguity in the position angles, the most recent (2021) observations by Tokovinin et al. (2022) were reduced with a methodology that is able to distinguish the correct quadrant.All other position angles were then changed as needed.After consulting the original sources, minor adjustments were made also to some of the uncertainties in order to make them more realistic, as they often account only for the internal errors.Observations published with no indication of their precision were assigned initial errors of 1 • in position angle (θ) and 2 mas in the separation (ρ).All uncertainties were later adjusted during the analysis, as we describe below.Three WDS measurements from 1991.9023, 2014.8568, and 2005.8688 were found to give abnormally large residuals, and were excluded.Table 2 gives the final list of speckle observations as used here.

ORBITAL ANALYSIS
A joint analysis of the CfA and Griffin velocities, and of the speckle measurements, was performed within a Markov chain Monte Carlo (MCMC) framework using the emcee package 2 of Foreman- Mackey et al. (2013).All position angles were adjusted for precession to the year J2000.We solved for the orbital period (P ), the angular semimajor axis (a ′′ ), the cosine of the inclination angle (cos i), the position angle of the ascending node for J2000.0 (Ω), the eccentricity (e) and argument of periastron for the primary (ω 1 ), cast as √ e cos ω 1 and 2 https://emcee.readthedocs.io/en/stable/index.html

√
e sin ω 1 , a reference time of periastron passage (T peri ), and the spectroscopic parameters K 1 , K 2 , and γ, which are the velocity semiamplitudes and center-of-mass velocity.An additional free parameter ∆ G was included to allow for a possible systematic offset between the CfA and Griffin velocities.It corresponds to the correction to be added to the Griffin RVs in order to place them on the CfA scale.To ensure proper weighting of the observations, six additional parameters were included in the analysis to represent multiplicative scaling factors for the formal uncertainties, which are not always accurate: f θ and f ρ for the astrometric measurements, f C,1 and f C,2 for the primary and secondary velocities from CfA, and f G,1 and f G,2 for the Griffin RVs.The total number of free parameters was 17.
We used 100 random walkers, and the MCMC chains had 10,000 links each, after burn-in.Convergence was checked by visual examination of the chains, and by requiring a Gelman-Rubin statistic of 1.05 or smaller (Gelman & Rubin 1992).Priors were all uniform over suitable ranges, except for those of the error scaling factors, which were log-uniform.
The results of the analysis are presented in Table 3. Derived properties are given at the bottom, and include the masses and the orbital parallax, π orb .A graphical representation of the spectroscopic orbit is shown in Figure 1, and the speckle orbit can be seen in Figure 2.
In addition to the present orbital solution and that of Griffin (2012), previous ones for vB 120 include a preliminary astrometric analysis of the Hipparcos data by Söderhjelm (1999), which assumed a circular orbit, an astrometric-spectroscopic analysis somewhat similar to ours by Tokovinin et al. (2015), in which they reported only the astrometric elements, and an astrometric so- -The values listed correspond to the mode of the posterior distributions, with uncertainties representing the 68.3% credible intervals.Priors in square brackets are uniform over the ranges specified, except those for the error inflation factors f , which are log-uniform.lution by Docobo et al. (2018), in which several of the elements were held fixed from the work of Griffin (2012).Table 4 presents a comparison of all of these results.

THE SPECTRAL ENERGY DISTRIBUTION (SED)
vB 120 has been observed in a variety of standard photometric systems.These measurements of the combined light can be used to derive estimates of the effective temperatures and angular diameters of the components, by comparison with appropriate synthetic spectra based on model atmospheres.Scaling the angular diameters by the distance derived from the orbital parallax obtained in the previous section then provides the absolute radii.We collected a total of 44 individual brightness measurements from the VizieR database 3 , in the following photometric systems: Johnson (U, B, V, R, I), Tycho-2 (B T , V T ), Hipparcos (Hp), Gaia DR3 (G, G BP , G RP ), Geneva (U, B1, B, B2, V 1, V, G), Pan-STARRS (g, r, z), the WBVR system (W, B, V, R), Strömgren (u, v, b, y), WISE (W 1-W 4), GALEX (N U V ), and 2MASS (J, H, K S ).Together, these observations span the entire optical range, and extend also into the UV and infrared (0.25-22 µm).
To further constrain the SED fit, we made use of the available estimates of the magnitude difference between the stars at different wavelengths.From our own spectroscopic observations in Section 2, we inferred ∆m = 0.44± 0.04 mag in the 5187 Å region covered by our spectra.A further estimate of ∆m ≈ 0.37 mag was reported by Griffin (2012), approximately in the V band.We assign it an uncertainty of 0.02 mag.Other measurements have been obtained by the speckle observers.The weighted average of those estimates in V is 0.37 ± 0.11 mag, and a measurement in the K band gave ∆m = 0.25±0.03mag.
As the SED is insensitive to the stars' surface gravities, we adopted typical values for dwarfs of log g = 4.5.The 3 https://vizier.cds.unistra.fr/viz-bin/VizieRFigure 3. Spectral energy distribution for vB 120, fitted using Kurucz stellar atmosphere models (Castelli & Kurucz 2003).The horizontal error bars represent the width of each bandpass.Residuals at the bottom are shown in units of magnitudes.The additional photometric error required by our analysis to achieve a reduced χ 2 value of unity is 0.025 ± 0.004 mag.The inset shows the predicted magnitude difference between the components from the best-fit model, along with the available ∆m measurements described in the text.
metallicity was held fixed at a value appropriate for the Hyades ([Fe/H] = +0.18;Dutra-Ferreira et al. 2016, see below).In addition to the individual temperatures and angular diameters, we solved for an additional photometric error added in quadrature to the published uncertainties of all the measured magnitudes, to account for possible underestimates of the observational errors as well as uncertainties in the various photometric zeropoints and calibrations.We also allowed the reddening E(B − V ) to be free, even though it is typically considered to be negligible for the Hyades (see, e.g., Taylor 2006).
We illustrate our fit in Figure 3.We obtained temperatures of 5656 ± 56 K and 5489 ± 60 K for the primary and secondary, where we have conservatively added 50 K in quadrature to the internal uncertainties to allow for the possibility of systematic errors in the models.The resulting angular diameters are 0.1969 ± 0.0020 mas and 0.1786 ± 0.0025 mas, leading to corresponding absolute radii of 0.968 ± 0.012 R ⊙ and 0.878 ± 0.013 R ⊙ .The radius uncertainties include the contribution from the formal error in the orbital parallax.The temperatures are slightly lower than our spectroscopic values by about 70 K, but are formally more precise.As expected, the reddening we infer is consistent with zero: E(B − V ) = 0.0047 ± 0.0053 mag.

STELLAR ACTIVITY
vB 120 has long been listed as a suspected variable star in the literature, with the designation NSV 1735.However, as pointed out by Griffin (2012), this seems to have been a consequence of a possible misprint in a single measurement of the V magnitude in the 1950s (V = 7.34), which had it about 0.4 mag brighter than all other determinations.No other reliable measurement as bright as this has been published since.
As it turns out, the very high photometric precision, Note.
-The elements held fixed in the Docobo solution were taken from the work of Griffin (2012).
high cadence, and continuity now attainable with spacebased missions such as Kepler/K2 and TESS have shown that most stars are variable at some level, and vB 120 is no exception.Douglas et al. (2019) used data from the K2 mission, and inferred a photometric period of 8.61 d, presumably due to rotation.Green et al. (2023) examined the TESS observations, and reported a period of 4.41 d, about half as long.
Figure 4 shows the measurements from TESS for the four sky sectors currently available (5, 32, 43, and 44), in which the peak-to-peak variation over this interval is only about 1.5 mmag.vB 120 is the brightest object in the photometric aperture.Sector 32 clearly shows a single dominant period of about 8.5 d, consistent with the estimate of Douglas et al. (2019).Sector 5, the one on which Green et al. (2023) based their result, as well as sectors 43 and 44, display a more complicated structure.Given the similarity of the properties of the two components, including their brightness, it seems plausible that both stars contribute to the variability with similar periods of roughly 8.5 d.In that case, it is possible that the corresponding peaks and troughs in the lightcurve happened to line up for the two stars in sector 32 (or the spots on one star disappeared), but were out of phase in the other sectors.We speculate that this misalignment could have led to the shorter period reported by Green et al. (2023) from sector 5.Alternatively, perhaps only one star is spotted, and its surface features come and go, or has two spots on opposite sides, one of which comes and goes.
If the rotation periods are both indeed ∼8.5 d, the projected rotational velocities would be expected to be ∼5.5 sin i rot km s −1 for an average radius of 0.9 R ⊙ , or about 5 km s −1 if the spin axes are parallel to the orbital axis (i rot ≈ i = 113.• 25).This value is not far from the adopted 6 km s −1 line broadening in our spectroscopic analysis of Section 2. It is also close to the v sin i estimate of Griffin (2012) for the primary (4.6 ± 0.7 km s −1 ), although his estimate for the secondary is much lower (2.1 ± 0.9 km s −1 ).A shorter rotation period of 4.41 d (Green et al. 2023) would require stronger spin/orbit misalignments to match the measured line broadening.
vB 120 is a known X-ray source, having been detected by the ROSAT and XMM-Newton missions.The strength of its chromospheric activity as measured by the emission cores in the Ca II H and K lines has been reported as log R ′ HK = −4.52 (Isaacson & Fischer 2010, average of three measurements) or log R ′ HK = −4.35 a Data downloaded from the Mikulski Archive for Space Telescopes, https://archive.stsci.edu/(Brown et al. 2022).While higher than the Sun, this level of activity is typical of Hyades stars with the spectral type of vB 120.Based on a similar log R ′ HK measure, Fuhrmeister et al. (2022) predicted a rotation period of 9.2 d using a statistical relation by Mittag et al. (2018), which is rather close to the period seen in the TESS data.

DISCUSSION
The absolute masses of vB 120, and its other physical properties, offer an opportunity for a comparison against current stellar evolution models at the age and composition of the Hyades.Both of these cluster attributes have some degree of uncertainty, however.For the age, a commonly quoted value is that of Perryman et al. (1998), 625 ± 50 Myr, obtained by isochrone modeling of the color-magnitude diagram (CMD) using the Hipparcos parallaxes.A very similar estimate of 650 ± 70 Myr, based on the lithium depletion boundary, was reported by Martín et al. (2018).It has been found that rotation can alter the shape of the upper main-sequence in the CMD, thereby affecting cluster ages.Accounting for this, Brandt & Huang (2015a) derived a rather older age for the Hyades of about 800 Myr, subsequently revised to 750 ± 100 Myr (Brandt & Huang 2015b).A similar analysis by Gossage et al. (2018) yielded an age of 680 Myr from the optical CMD, which they favored over a larger estimate of about 740 Myr in the near infrared.For a compilation of other age estimates, see Douglas et al. (2019).
Metallicity determinations for the Hyades have typically ranged between [Fe/H] = +0.1 and +0.2 dex.Classical spectroscopic studies of FGK-type dwarfs find a metallicity of about +0.13 or +0.14 dex (e.g., Cayrel et al. 1985;Boesgaard & Friel 1990;Paulson et al. 2003;Schuler et al. 2006).Estimates for the giants give similar values (Smith 1999;Carrera & Pancino 2011;Ramya et al. 2019).Variations from star to star have been reported, particularly among the A stars, as well as systematic differences between Fe I and Fe II abundances, increasing toward the cooler stars (e.g., Aleo et al. 2017).Dutra-Ferreira et al. (2016) examined the importance of the line list, the consistency between dwarfs and giants, and different ways of constraining other stellar parameters needed for the analysis (T eff , log g, microturbulent velocity).They expressed a preference for the value [Fe/H] = +0.18± 0.03, which they found to be the same for dwarfs and giants, and to be robust against the choice of the line list.
For the purposes of this paper, we adopt a Hyades metallicity of [Fe/H] = +0.18from Dutra-Ferreira et al. (2016), and an age of 750 Myr from Brandt & Huang (2015b).The impact of these choices is discussed below.
Figure 5 shows the empirical mass-luminosity relation for the Hyades in the visual band.It includes the components of all binary systems with previously measured dynamical masses.vB 120 is in general agreement with the trend (slope), although both components fall slightly below the stellar evolution models shown in the figure for the adopted age and metallicity of the Hyades, by about 1 or 1.5σ (see the inset).Other stars between 1 and 1.5 M ⊙ also tend to be slightly fainter than predicted by theory, on average.One possible explanation, though it seems unlikely, might be systematic errors affecting the magnitudes and/or masses of several of these systems in the same way.Another could be missing opacities in the stellar atmosphere models used to compute the fluxes for the isochrones.A similar diagram for the K band (Figure 6) indicates that the primary of vB 120 is consistent with both the MIST model isochrone of Choi et al. (2016) and the PARSEC v1.2S model of Chen et al. (2014), while the secondary is marginally below the latter.The one other system in the Hyades for which the individual masses and K-band absolute magnitudes are known is HD 284163 (Torres et al. 2023), a system of three stars also shown in the figure .A comparison between those same models and the components' effective temperatures and radii, from our SED analysis of Section 5, is shown in Figure 7.The radii are consistent with the models, within their uncertainties.The effective temperatures agree well with the MIST isochrone, and are just slightly cooler but essentially also within 1σ of the PARSEC v1.2S models, which are in turn systematically hotter than MIST by about 70 K.
Adopting a different age and/or composition for the Hyades changes the predicted absolute magnitudes from the models by a small but non-negligible amount that is mass-dependent.Figure 8 quantifies this for the PARSEC v1.2S models in the V band, addressing the comparison shown previously in Figure 5.For exam-   ple, lowering the metallicity from our chosen value of [Fe/H] = +0.18 to +0.13 dex at a fixed age (dashed line in Figure 8) makes the models slightly brighter (negative ∆M V ).The effect is roughly 0.04 mag for masses larger than about 1.2 M ⊙ , gradually increasing toward lower masses.On the other hand, reducing the age from 750 Myr to 625 Myr at a fixed metallicity makes the predicted magnitudes fainter (positive ∆M V ), by up to 0.04 mag for unevolved stars with masses below about 1.7 M ⊙ .For higher masses the differences are larger, as such stars begin to evolve (solid line).Overall, we estimate that the uncertainty in the age and composition of the Hyades can affect the comparison in Figure 5 at a level similar to the observational errors in M V .
The orbital parallax of vB 120 from our analysis, π orb = 21.86 ± 0.15 mas, has a formal error of only 0.7%.The entry for object in the Gaia DR3 catalog (source identifier 3404812685132622592; Gaia Collaboration et al. 2022) is lower (20.53 ± 0.18 mas)4 , but did not account for the orbital motion.A sign of this is evident in the reported quality of the astrometric fit, as measured by the Renormalized Unit Weight Error (RUWE).For well-behaved sources in which a single-star model provides a good fit, the RUWE is expected to be near 1.0.Values larger than about 1.4 could be indicative of a non-single source, or an otherwise problematic astrometric solution.vB 120 has a RUWE of 7.116.While this does not necessarily imply the parallax is biased, at the very least the formal uncertainty will be underestimated.Following the prescription by Maíz Apellániz (2022), we estimate the external error of the Gaia DR3 parallax to be 0.63 mas.Even with this larger error, the difference compared to our orbital parallax is still at the 2σ level, likely a consequence of unmodeled binary motion in Gaia.
The Hipparcos mission delivered a parallax for vB 120 of 23.64 ± 0.99 mas (source identifier HIP 22505; van Leeuwen 2007), which is about 2σ larger than ours.This also did not account for orbital motion.Madsen et al. (2002) applied the moving-cluster method to the Hyades using Hipparcos positions and proper motions, and obtained purely kinematic parallaxes for many of the cluster members that are often better than the trigonometric values.Their result for vB 120 is 21.85 ± 0.39 mas, which is rather different from the trigonometric value from the mission, but is nearly identical to ours.The study also derived an astrometric radial velocity for vB 120 of 41.81 ± 0.60 km s −1 , again very close to the γ velocity listed in our Table 3.A similar study using astrometry from the Gaia DR1 catalog (Gaia Collaboration et al. 2016) was carried out by Reino et al. (2018), who reported a kinematic parallax of 21.58 ± 0.23 mas, along with an astrometric RV of 41.97 ± 1.05 km s −1 .Both are consistent with the results of this paper.

CONCLUSIONS
In this work we have made use of astrometric as well as new and existing spectroscopic observations, to present improved dynamical mass estimates for another binary system in the Hyades, vB 120.This is the ninth such example in the cluster.With formal mass uncertainties for the primary and secondary under 1.8%, it ranks among the best determinations in this small group.We also derived the orbital parallax of the system to better than 0.7%, and presented evidence of its superior accuracy (and precision) compared to Gaia DR3, based on independent kinematic parallax determinations by others.
These observational results show relatively good agreement in the mass-luminosity diagram with two sets of current stellar evolution models (MIST, and PAR-SEC v1.2S) for the known age and metallicity of the clus-ter, in both the V and K bandpasses.Only one other Hyades system is available for this type of comparison in the K band (HD 284163;Torres et al. 2023).
Even though vB 120 is not eclipsing, we have also obtained estimates of the absolute radii and effective temperatures of the components through a spectral energy distribution analysis, supplemented by the orbital parallax.Those results are also consistent with theory, with the temperatures slightly favoring the MIST models.
Examination of the TESS photometry for vB 120 shows variability at a low level (≤ 1.5 mmag total amplitude), with a period of about 8.5 d that is similar to a previous estimate from the literature, and is presumably due to rotation combined with variable surface activity on one or both stars.

Figure 1 .
Figure 1.Radial-velocity measurements for vB 120 from the CfA and Griffin (2012), along with our adopted model for the spectroscopic orbit.The dotted line at the top represents the centerof-mass velocity.Residuals are shown at the bottom.Phase 0.0 corresponds to periastron passage.

Figure 2 .
Figure2.Speckle observations for vB 120 with our adopted model for the astrometric orbit.The "+" sign represents the primary star.Short line segments connect the measured position with the predicted location of the secondary on the orbit.The line of nodes is indicated with a dotted line, and Ω marks the ascending node.The location of periastron is indicated with the square marked "P".

Figure 4 .
Figure 4. TESS light curve of vB 120 in the four sectors observed by the satellite as of this writing.a .Fluxes in each sector have been normalized by the median value.a Data downloaded from the Mikulski Archive for Space Telescopes, https://archive.stsci.edu/

Figure 5 .
Figure 5. Mass-luminosity relation for the Hyades cluster in the visual band.The masses for vB 120 are from Table 3, and the absolute magnitudes are based on a typical system brightness of V = 7.73 ± 0.02, a V -band magnitude difference of 0.37 ± 0.02 (Griffin 2012), the orbital parallax from the present work, and the assumption of zero extinction.Measurements for the other systems are taken from Torres (2019), with updates for two of them by Brogaard et al. (2021) and Anguita-Aguero et al. (2022), and from Torres et al. (2023).For comparison, model isochrones from two different series of calculations are also shown, for the age (750 Myr; Brandt & Huang 2015b) and metallicity of the cluster.The inset shows a close-up of vB 120.

Figure 6 .
Figure 6.Similar to Figure 5, for the K band.The individual absolute magnitudes for vB 120 use the system brightness from 2MASS, and the magnitude difference of 0.25 ± 0.03 from Patience et al. (1998).HD 284163 (also shown) is the only other multiple system in the Hyades with known masses and individual K-band (Torres et al. 2023).

Figure 7 .
Figure 7. Individual radii and effective temperatures for the components of vB 120 from our SED fit, compared with the same stellar evolution models as in Figures 5 and 6.

Figure 8 .
Figure 8. Influence of the age and metallicity on the absolute Vband magnitude predicted by the PARSEC v1.2S models for the Hyades, as a function of mass.The dashed line shows changes in M V for different compositions, at a fixed age of 750 Myr (our reference age).The solid line shows changes for different ages at a fixed reference composition of [Fe/H] = +0.18.Differences between models are taken in the sense indicated in the labels.

Table 2
Speckle Measurements for vB 120

Table 3
Orbital Parameters for vB 120

Table 4
Orbital Parameters for vB 120 from this Work Compared with Previous Determinations