K-dwarf Radius Inflation and a 10-Gyr Spin-down Clock Unveiled through Asteroseismology of HD 219134 from the Keck Planet Finder Yaguang Li (李亚光), Daniel Huber, J. M. Joel Ong (王加冕), Jennifer van Saders, R.R. Costa, Jens Reersted Larsen, Sarbani Basu, Timothy R. Bedding, Fei Dai (戴飞), Ashley Chontos, Theron W. Carmichael, Daniel Hey, Hans Kjeldsen, Marc Hon, Tiago L. Campante, Mário J. P. F. G. Monteiro, Mia Sloth Lundkvist, Nicholas Saunders, Howard Isaacson, Andrew W. Howard, Steven R. Gibson, Samuel Halverson, Kodi Rider, Arpita Roy, Ashley D. Baker, Jerry Edelstein, Chris Smith, Benjamin J. Fulton, Josh Walawender Abstract We present the first asteroseismic analysis of the K3 V planet host HD 219134, based on four consecutive nights of radial velocities collected with the Keck Planet Finder. We apply Gold deconvolution to the power spectrum to disentangle modes from sidelobes in the spectral window, and extracted 25 mode frequencies with spherical degrees 0 ≤ ℓ ≤ 3 0\leq\ell\leq 3 . We derive the fundamental properties using five different evolutionary-modeling pipelines and report a mass of 0.763 ± \pm 0.020 (stat) ± \pm 0.007 (sys) M ⊙ \text{M}_{\odot} , a radius of 0.748 ± \pm 0.007 (stat) ± \pm 0.002 (sys) R ⊙ \text{R}_{\odot} , and an age of 10.151 ± \pm 1.520 (stat) ± \pm 0.810 (sys) Gyr. Compared to the interferometric radius 0.783 ± \pm 0.005 R ⊙ \text{R}_{\odot} , the asteroseismic radius is 4% smaller at the 4- σ \sigma level — a discrepancy not easily explained by known interferometric systematics, modeling assumptions on atmospheric boundary conditions and mixing lengths, magnetic fields, or tidal heating. HD 219134 is the first main-sequence star cooler than 5000 K with an asteroseismic age estimate and will serve as a critical calibration point for stellar spin-down relations. We show that existing calibrated prescriptions for angular momentum loss, incorporating weakened magnetic braking with asteroseismically constrained stellar parameters, accurately reproduce the observed rotation period. Additionally, we revised the masses and radii of the super-Earths in the system, which support their having Earth-like compositions. Finally, we confirm that the oscillation amplitude in radial velocity scales as ( L / M ) 1.5 (L/M)^{1.5} in K dwarfs, in contrast to the ( L / M ) 0.7 (L/M)^{0.7} relation observed in G dwarfs. These findings provide significant insights into the structure and angular momentum loss of K-type stars. show]yaguangl@hawaii.edu I Introduction Asteroseismology, the study of stellar oscillations, is a powerful method for measuring fundamental stellar properties and studying interior dynamics. Early detections of oscillations in solar-type stars—see T. R. Bedding & H. Kjeldsen [14] for a review—were made possible through ground-based radial velocity (RV) observations, such as those observed in Procyon [ 23 , 139 , 124 ] , \beta Hyi [ 16 ] and \alpha Cen A [ 20 , 21 ] . These observations confirmed the presence of solar-like oscillations in cool stars beyond the Sun, which are stochastically excited and damped by surface turbulence [ 71 ] . In stars with solar-like oscillations, acoustic modes are observed around the frequency of maximum power, \nu_{\text{max}} . Their non-rotating frequencies are well approximated by the asymptotic relation involving two quantum numbers, the radial order, n , and angular degree, \ell [ 177 , 165 ] : \nu_{n,\ell}\simeq\Delta\nu\left(n+\frac{\ell}{2}+\epsilon\right)-\delta\nu_{0,\ell}. (1) The large frequency separation, \Delta\nu , depends primarily on the sound travel time across the star and, to first order, scales with the square root of the mean stellar density [ 186 ] . The phase offset, \epsilon , generally falls between 0.8 and 1.6 in main-sequence stars [ 196 ] , and depends on the physical structure of the star, primarily as concentrated near the center and surface [ 161 ] . The small separations, \delta\nu_{0,\ell} , quantify the frequency offsets between modes of different \ell but with same n . In main-sequence stars, \delta\nu_{0,2} is sensitive to the chemical gradient near the core [ 178 , 160 ] , making it a useful diagnostic for tracing hydrogen-burning processes and, therefore, for determining stellar ages on the main sequence [ 38 , 195 , 86 ] . Since the advent of space missions in the 2000s, photometric measurements have become a primary method for detecting stellar oscillations, with the significant advantage of allowing simultaneous observations of many stars. Missions like WIRE [ 24 ] , MOST [ 125 ] , CoRoT [ 8 ] , Kepler [ 19 ] and TESS [ 156 ] have substantially increased the number of detections of solar-like oscillations in G- and F-type main-sequence stars, as well as in sub- and red giants spanning a wide range of evolutionary stages [ 33 , 97 , see reviews by] . The high precision of these photometric observations allows for detailed analysis of oscillation modes, probing the structure and dynamics of stellar interiors, including rotation, magnetism, and binary merging [see 2 , for a review] . Recent significant advances in Doppler precision, driven by the search for Earth-like planets, have opened new opportunities for detecting stellar oscillations with much lower amplitudes. Extreme Precision Radial Velocity (EPRV) techniques now achieve radial velocity measurements with precisions in the range of 10–30 cm/s [ 60 , 200 ] . This level of precision makes it possible to detect the intrinsic oscillations of K dwarfs, which have amplitudes ranging from 1 to 6 cm/s. These oscillations are very challenging to detect in photometry due to the presence of granulation noise, which is significantly lower (relative to the oscillations) in RV measurements from spatially-averaged line profiles [ 79 , 76 , 7 , 103 ] . Until recently, the only K dwarfs with oscillations detected in RV were the bright K1 V star \alpha Cen B [ 28 , 105 ] and the slightly hotter K0 V star 70 Oph A [ 29 ] . The arrival of EPRV instruments has ushered in a new era of K-dwarf asteroseismology Oscillations in the K5 V star \epsilon Indi were reported by T. L. Campante et al. [26] with six half-nights of observations with ESPRESSO on the VLT [ 153 ] , and also at lower signal-to-noise with HARPS and UVES by M. S. Lundkvist et al. [120] . M. Hon et al. [87] detected oscillations in the K0 V star \sigma Draconis using one night of observations with the Keck Planet Finder (KPF) on the Keck I telescope [ 65 , 66 , 67 , 68 ] . Although TESS photometryalso shows oscillations in \sigma Draconis from 14 sectors of data, the KPF RV power spectrum revealed oscillation modes over a broader frequency range due to the higher signal-to-noise ratio achieved in the RV data. Figure 1 illustrates the expanding asteroseismology footprint, with EPRV techniques populating the K dwarf parameter space. While a single night of data is sufficient to detect both \nu_{\text{max}} and \Delta\nu (provided sufficient SNR is reached), achieving the resolution required to measure small frequency spacings like \delta\nu_{0,2} remains challenging. For example, an 8-hour time series only provides a frequency resolution of \approx 34 \mu{\rm Hz} , which exceeds the typical \delta\nu_{0,2} values of about 10 \mu{\rm Hz} for main-sequence stars [ 195 , e.g.] . We can address this limitation using multiple nights of data. For this paper, we have obtained four nights of data collected with the KPF, corresponding to a frequency resolution of \approx 3 \mu{\rm Hz} . We report the detection of oscillations in HD 219134 and the determination of its benchmark mass, radius, and age via asteroseismology. I.1 The K3 V exoplanet host HD 219134 HD 219134 (HR 8832) is a nearby K3 V dwarf ( d=6.5 pc; V=5.57 ). It has an angular diameter precisely measured through interferometry [ 112 , 91 , 57 ] , enabling an accurate determination of its radius (see § III.1 ). Asteroseismology provides an independent method to estimate the stellar radius by probing the star’s interior through oscillation frequencies. This star presents an opportunity to directly compare these two techniques. HD 219134 has a rotation period of \approx 42 days [ 140 , 61 ] . Cool stars experience gradual spin-down over time, making rotation period a valuable proxy for stellar age — the principle underpinning gyrochronology. However, the applicability of the rotation-age relation relies on calibrations using stars with well-known ages, primarily those in open clusters [ 45 ] . Importantly, HD 219134 rotates more slowly than all stars of similar spectral type in the current open cluster calibrators. Furthermore, recent studies of K-dwarf stars have revealed a phase of apparently stalled spin-down [ 44 , 45 , 3 ] . It remains unclear for how long the epoch of stalling persists or the degree to which it affects the rotation rates of old stars. Measuring the age of this star using asteroseismology would provide a valuable test of the rotation-age spin-down relationship for stars at advanced ages. HD 219134 hosts five known RV planets, including the transiting super-Earths HD 219134 b and c, whose orbital periods are 3.09 and 6.87 days [ 193 , 140 , 166 , 158 , 69 , 106 ] . A detailed seismic characterization of the host star’s mass and radius would refine the properties of these planets. The paper is organized as follows. We report our detection of oscillations and extract oscillation frequencies (§ II ). We derive stellar properties using asteroseismic modeling (§ III ). We highlight the discrepancy between our asteroseismic radius and an interferometric radius and explore potential explanations (§ IV ). We discuss the implication of stellar age for gyrochronology by testing models with angular momentum loss (§ V ). We revise planet properties using the new stellar mass and radius (§ VI ). Finally, we analyze mode amplitudes (§ VII ), and present our conclusions (§ VIII ). Figure 1: H–R diagram showing the footprint of asteroseismology conducted using space-based photometry [ 34 , 25 , 117 , 118 , 168 , 201 , 109 , 80 , 73 , 203 , 119 , e.g.] and ground-based radial velocity (RV) measurements [ 105 , 29 , 180 , e.g.] . The diagram highlights recent contributions made possible by Extreme Precision Radial Velocity (EPRV) instruments [ 26 , 87 ] , which have significantly expanded the parameter space for K dwarf studies. Figure 2: Asteroseismic observations of HD 219134 over four consecutive nights using the Keck Planet Finder. Top: Radial-velocity time series after filtering out signals with periods longer than 1.2 hours. Bottom: Power spectrum of the RV time series, weighted by the reported RV uncertainties, displaying a clear power excess around 4500 \mu{\rm Hz} . The inset shows the spectral window. Figure 3: Échelle diagrams showing structures of regular frequency spacings. Panel a: replicated échelle diagram of the original power spectrum. Panel b: échelle diagram of the power spectrum, deconvolved from the spectral window and then convolved with a 1 \mu Hz width Gaussian filter for clarity. Panel c: collapsed échelle diagram, summing the power along the y-axis of panel b. Panel d: same as panel b, but highlighting the extracted oscillation modes. II Oscillation Frequencies Table 1: Oscillation frequencies of HD 219134, including corrections for line-of-sights Doppler shifts. n \ell \nu_{n,\ell} ( \mu{\rm Hz} ) \sigma(\nu_{n,\ell}) ( \mu{\rm Hz} ) 18 0 3561.03 0.58 19 0 3742.84 1.07 21 0 4106.45 0.64 23 0 4468.50 0.91 25 0 4836.80 0.95 26 0 5019.39 0.85 27 0 5201.19 0.95 30 0 5752.07 1.03 17 1 3465.84 0.84 18 1 3647.65 0.43 19 1 3830.23 0.90 20 1 4011.26 0.65 24 1 4739.26 0.95 25 1 4923.41 0.64 17 2 3549.33 0.62 20 2 4094.75 0.85 21 2 4276.55 1.05 23 2 4641.73 1.00 24 2 4825.09 0.52 26 2 5193.39 1.12 28 2 5559.34 0.98 16 3 3445.55 1.05 19 3 3992.53 1.07 23 3 4722.10 0.93 24 3 4907.80 1.02 We obtained approximately 9 hours of observations per night over 4 consecutive nights, from September 25 to 28, 2024 (HST), using the Keck Planet Finder (KPF) on the Keck I telescope [ 65 , 66 , 67 , 68 ] . A total of 2,155 radial velocity measurements were collected. We used 45-second exposures, with a readout time of 15 seconds, resulting in a cadence of 1 minute per observation. Observing conditions were excellent and stable, with seeing ranging from 0.4” to 0.8”. The signal-to-noise ratio (SNR) of the spectra at 747 nm remained consistently around 500. The spectra were processed with KPF Data Reduction Pipeline [ 67 , DRP;] 1 1 1 https://github.com/Keck-DataReductionPipelines/KPF-Pipeline?tab=readme-ov-file . KPF DRP executes several processing steps, including quadrant stitching, flat-field correction, order tracing, optimal spectral extraction, and wavelength calibration. RVs are determined from a cross-correlation function (CCF) mask method, using the same K2 binary CCF mask as in the ESPRESSO pipeline [ 153 ] . A mean position of the CCF is fitted using a Gaussian plus top-hat function. RVs are computed separately for each of the three KPF slices and for each camera (green and red). The RVs from individual slices are combined through a weighted average, where the weights are inversely proportional to the photon-limited RV uncertainties for each slice. The RVs from the cameras are adjusted by median subtraction and corrected for instrumental drift, based on laser frequency comb spectra obtained before and after the stellar RV observations. The drift-corrected RV time series for the two cameras are then combined into a single dataset using a flux-weighted mean. To calculate the power spectrum of the RV time series, we applied the Lomb-Scargle method [ 113 , 164 ] , using the reported RV uncertainties as weights [ 64 ] . Normalization of the power spectrum followed the convention of H. Kjeldsen & T. R. Bedding [102] : a sine wave with amplitude A produces a peak with a height of A^{2} in the power spectrum, and A^{2}T_{\rm obs} in the power density spectrum, with T_{\rm obs} being the effective observing duration. In Figure 2 , we show the RV time series, along with the power spectrum. The power spectrum in the high-frequency regime contains only white noise, from which we estimated the time-domain RV scatter, \sigma_{\rm noise} , due to photon and instrumental noise. The scatter was calculated as \sigma_{\rm noise}=\sqrt{P_{\rm mean}/(2\Delta t)} , where P_{\rm mean} is the mean power density between 6500 \mu{\rm Hz} and 8000 \mu{\rm Hz} , and \Delta t is the cadence. We obtained a value of time-domain scatter at 0.37 m/s, which is in agreement with the average RV uncertainties reported from the reduction pipeline at 0.41 m/s. Regular data gaps in time domain inevitably cause multiple peaks near the true frequency in the Fourier domain, a phenomenon known as spectral leakage. The observed power spectrum is the convolution of the true spectrum (as from a continuous and infinite time series) with the spectral window function. To examine the shape of the spectral window for our dataset, we generate sine and cosine waves at an arbitrary test frequency, sampled at the timestamps of our RV data. We then calculate and average the power spectra of both waves, which helps reduce edge effects from signal apodization. The inset of Figure 2 shows the spectral window centred around the test frequency. The strongest sidelobes appear at multiples of 1 c/d ( \approx 11.57 \mu{\rm Hz} ), due to the existence of daily gaps. The daily sidelobes present a significant challenge in identifying the \delta\nu_{0,2} spacing, as it is very close to 1 c/d in K-type main-sequence stars [ 195 ] . This can be seen from Figure 3 (a), which shows the power spectrum folded into segments of \Delta\nu and displayed in an échelle format. In such diagrams, modes of the same spherical degree \ell form approximately vertical ridges from consecutive radial orders n , as predicted by the asymptotic relation (Equation 1 ). Based on observations and theoretical predictions of the phase offset \epsilon [ 195 , 145 ] , we expect \epsilon to be approximately 1.4 given the \Delta\nu value of HD 219134. This implies that the ridges in Figure 3 (a) at an abscissa of \approx 90 \mu{\rm Hz} corresponds to the \ell=0 and \ell=2 modes, while the ridge at 180 \mu{\rm Hz} corresponds to the \ell=1 (and \ell=3 , if present) modes. Due to geometric cancellation, only oscillation modes with low spherical degrees ( 0\leq\ell\leq 3 ) are expected to be visible. However, Figure 3 (a) shows more than four vertical ridges, many of which are sidelobes — replicas of true ridges with offsets at multiples of 1 c/d. In particular, the true \ell=2 ridge could overlap with the left sidelobe of the \ell=0 ridge, due to the closeness of {\delta\nu_{0,2}} to 1 c/d. To better reveal the frequency structure, we deconvolved the power spectrum against the spectral window using the Gold algorithm [ 70 ] , by applying the numerical implementation described by M. Morháč et al. [138] . The algorithm takes the power spectrum and the spectral window as inputs, and iteratively solves for the underlying true spectrum with some regularization 2 2 2 https://github.com/parallelpro/maemae . We selected this algorithm over other deconvolution schemes for its desirable properties: (1) it ensures the solution remains positive when both the input data and spectral window are positive; and (2) it provides non-oscillating solutions for isolated peaks, effectively suppressing ringing artifacts [ 137 ] . A quantitative evaluation of its performance in asteroseismic applications will be presented in a forthcoming paper; in this work, we applied it solely as a visual aid to mode identification, but otherwise applied traditional fitting techniques to derive estimates and uncertainties of mode frequencies. Figure 3 (b) shows the spectrum after deconvolution, revealing clear \ell=0 – 3 mode ridges. These are more apparent in Figure 3 (c), which shows the collapsed power spectrum from summing power along the vertical direction in Figure 3 (b). From Figure 3 (b) and (c), we can identify the peaks along each ridge, which we preliminarily considered as candidate oscillation modes. We extracted their frequencies by simultaneously fitting sine waves at these frequencies to the RV time series, obtaining the corresponding amplitudes and phases. Signals with amplitudes exceeding 3.5 times the noise level (0.37 m/s) were retained as the final list of modes. These are highlighted in Figure 3 (d). Furthermore, we corrected the line-of-sight Doppler shifts for these frequencies following G. R. Davies et al. [49] , using a radial velocity of -18368 m/s based on an average value over four nights. This adjustment reduced the frequencies by an average of \approx 0.27 \mu{\rm Hz} . Frequency uncertainties were calculated using the method described by H. Kjeldsen & T. R. Bedding [104] : \sigma(\nu)\approx 0.44\sqrt{\pi/N}\sigma_{\rm noise}a^{-1}\sqrt{T_{\rm obs}^{-2}+\tau^{-2}}, (2) where N is the number of points in the time series, a is the mode amplitude, and \tau is the mode lifetime. For the Sun, the mode lifetime is typically 1–3 days [ 32 , 15 ] , and it increases for cooler stars. We assumed a conservative mode lifetime of 3 days for this star in the uncertainty calculation. Table 1 lists the determined oscillation frequencies. Furthermore, we fitted Equation 1 to these frequencies and obtained the following asymptotic parameters: {\Delta\nu}{}=182.799\pm 0.069 \mu{\rm Hz} , \epsilon=1.455\pm 0.010 , {\delta\nu_{0,1}}=3.81\pm 0.41 \mu{\rm Hz} , {\delta\nu_{0,2}}=10.90\pm 0.41 \mu{\rm Hz} , and {\delta\nu_{0,3}}=22.27\pm 0.59 \mu{\rm Hz} . III Asteroseismic Modeling In this section, we derive the stellar parameters of HD 219134 using modern asteroseismic modeling techniques. We first describe the input observables used in the modeling process (§ III.1 ), followed by an overview of the modeling methodology and key assumptions (§ III.2 ). Finally, we provide brief comments on the derived stellar age and its implications (§ III.3 ). III.1 Classical Constraints Figure 4: Stellar mass, radius and age derived from five independent asteroseismic modeling teams. The asteroseismic radius is compared against the interferometric radius based on an angular diameter measurement from A. Elliott et al. [57] . To derive a bolometric flux for HD 219134 we used optical broadband photometry from Tycho [ 85 ] , since near-infrared 2MASS photometry is saturated and systematic errors for bright stars are not yet well-calibrated in Gaia [ 157 ] . The star has a B_{\rm T} magnitude of 6.861\pm 0.015 and a V_{\rm T} magnitude of 5.674\pm 0.009 . Incorporating the Gaia DR3 distance and bolometric corrections, we used ISOCLASSIFY [ 93 , 17 ] to calculate bolometric fluxes of {{f_{\rm bol}}}=(1.881\pm 0.055)\times 10^{-7} erg s -1 cm -2 and {{f_{\rm bol}}}=(1.981\pm 0.054)\times 10^{-7} erg s -1 cm -2 , based on B_{\rm T} and V_{\rm T} , respectively. R. Ligi et al. [112] determined the bolometric flux by fitting its spectral energy distribution, finding {{f_{\rm bol}}}=(1.986\pm 0.024)\times 10^{-7} erg s -1 cm -2 . Due to the higher precision of the V_{\rm T} magnitude, we adopted its corresponding bolometric flux value. To refine the uncertainty, we calculated the standard deviation of all three flux measurements to account for systematic uncertainties, and combined this in quadrature with the original uncertainty. This yielded a final value of {{f_{\rm bol}}}=(1.981\pm 0.072)\times 10^{-7} erg s -1 cm -2 . The star’s angular diameter was measured using CHARA/VEGA by R. Ligi et al. [112] as \theta=1.035\pm 0.021 mas, after correcting for limb darkening. Combined with the Gaia DR3 distance of D=6.541\pm 0.002 pc [ 10 ] , this results in a stellar radius of R=0.728\pm 0.015 \text{R}_{\odot} via R=\frac{1}{2}\theta D. (3) We calculated the star’s interferometric effective temperature based on the bolometric flux and angular diameter: {T_{\text{eff}}}=4^{1/4}f_{\rm bol}^{1/4}\sigma_{\rm SB}^{-1/4}\theta^{-1/2}, (4) where \sigma_{\rm SB} is the Stefan-Boltzmann constant. The result is {T_{\text{eff}}}{}=4854\pm 66 K, consistent with the spectroscopic temperature measured by L. J. Rosenthal et al. [158] , who reported {T_{\text{eff}}}=4817\pm 62 K. The measured metallicity from L. J. Rosenthal et al. [158] is {{\rm[Fe/H]}}{}=0.083\pm 0.058 dex. A recent measurement of the star’s angular diameter by A. Elliott et al. [57] is \theta=1.114\pm 0.007 mas, which corresponds to a stellar radius of R_{\star}=0.783\pm 0.015 \text{R}_{\odot} , and an interferometric temperature of {T_{\text{eff}}}{}=4678\pm 45 K. In principle, this angular diameter should be more accurate, given that the measurement spanned a wider range of spatial frequencies. However, asteroseismic modeling using this interferometric T_{\text{eff}} predicts a radius that differs from the interferometric radius by >3\sigma . This suggests the two methods may not be strictly compatible, resulting from either unaccounted interferometric systematics or inaccuracies in the stellar models (see § IV for detail). To maximise compatibility between inputs and stellar models, we adopted a conservative set of input observables for modeling: the interferometric T_{\text{eff}} based on R. Ligi et al. [112] , spectroscopic [Fe/H] from L. J. Rosenthal et al. [158] , and the oscillation frequencies presented in Table 1 . We deliberately excluded the bolometric flux and radius as direct constraints, despite their high precision, to avoid potential inconsistencies between these values and the asteroseismic constraints. We summarize the stellar properties in Table 3 . Table 2: Summary of the frequency modeling. Team 1 Team 2 Team 3 Team 4 Team 5 Model Configuration Stellar evolution code MESA r240301 MESA r12778 GARSTEC v2015 YREC GARSTEC v2020 Pulsation code GYRE v7.1 GYRE v5 ADIPLS v0.4 H. M. Antia & S. Basu [5] GYRE v7.2 Opacities & EOS OPAL OPAL OPAL OPAL OPAL Atmospheric boundary condition P. H. Hauschildt et al. [81] , P. H. Hauschildt et al. [82] F. Castelli & R. L. Kurucz [31] K. S. Krishna Swamy [107] A. S. Eddington [56] A. S. Eddington [56] R. Trampedach et al. [185] Metal mixture M. Asplund et al. [6] M. Asplund et al. [6] M. Asplund et al. [6] N. Grevesse & A. J. Sauval [74] N. Grevesse & A. J. Sauval [74] Nuclear reaction rate R. H. Cyburt et al. [46] C. Angulo et al. [4] R. H. Cyburt et al. [46] C. Angulo et al. [4] C. Angulo et al. [4] A. Formicola et al. [63] J. W. Hammer et al. [78] E. G. Adelberger et al. [1] A. Formicola et al. [63] C. Angulo et al. [4] A. Formicola et al. [63] J. W. Hammer et al. [78] Mixing length formulation L. Henyey et al. [83] J. P. Cox & R. T. Giuli [42] E. Böhm-Vitense [18] E. Böhm-Vitense [18] J. P. Cox & R. T. Giuli [42] Mixing length parameter {\alpha_{\rm MLT}} Free Solar-calibrated Free Free Varying vs. solar-calibrated \alpha_{{\rm MLT},\odot} 1.96 1.71 1.79 1.84 1.83 Convective overshoot F. Herwig [84] f_{\rm ov,shell}=0.0174 None None None None Extra-mixing None Gravitational settling Turbulent mixing A. A. Thoul et al. [181] A. A. Thoul et al. [181] A. A. Thoul et al. [181] Diffusive mixing Surface correction W. H. Ball & L. Gizon [12] two-term + Y. Li et al. [110] ensemble correction T. Sonoi et al. [173] W. H. Ball & L. Gizon [12] cubic-term W. H. Ball & L. Gizon [12] two-term W. H. Ball & L. Gizon [12] two-term + I. W. Roxburgh [159] \epsilon -matching Grid free parameters M , [M/H], \alpha_{\rm MLT} , Y_{\rm init} M , [M/H], Y_{\rm init} M , [M/H], \alpha_{\rm MLT} , Y_{\rm init} M , [M/H], \alpha_{\rm MLT} , Y_{\rm init} M , [M/H], Y_{\rm init} Derived Stellar Parameters Mass M_{\star} ( \text{M}_{\odot} ) 0.763\pm 0.020 0.779\pm 0.031 0.773\pm 0.038 0.776\pm 0.011 0.779\pm 0.019 Radius R_{\star} ( \text{R}_{\odot} ) 0.748\pm 0.007 0.753\pm 0.010 0.751\pm 0.012 0.750\pm 0.003 0.754\pm 0.006 Age t_{\star} (Gyr) 10.2\pm 1.5 8.9\pm 2.0 9.2\pm 2.5 8.3\pm 0.6 8.1\pm 1.3 Density \rho_{\star} ( \rho_{\odot} ) 1.823\pm 0.003 1.825\pm 0.001 1.826\pm 0.100 1.824\pm 0.001 1.818\pm 0.007 Surface gravity {\log g}_{\star} (dex) 4.577\pm 0.001 4.576\pm 0.006 4.570\pm 0.005 4.575\pm 0.002 4.575\pm 0.004 Y_{\rm init} 0.270\pm 0.023 0.278\pm 0.022 0.279\pm 0.019 0.280\pm 0.005 0.272\pm 0.010 {\alpha_{\rm MLT}}{}/{\alpha_{\rm MLT}}_{,\odot} 0.95\pm 0.05 — 1.03\pm 0.07 0.94\pm 0.04 — Table 3: Star and planet properties. Property Value Reference HD 219134 Photometry B_{T} -band magnitude 6.861 \pm 0.015 E. Høg et al. [85] V_{T} -band magnitude 5.674 \pm 0.009 E. Høg et al. [85] Bolometric flux {f_{\rm bol}} (erg s -1 cm -2 ) (1.981 \pm 0.072) \times 10^{-7} This work Astrometry Distance (pc) 6.5409 \pm 0.0023 C. A. L. Bailer-Jones et al. [11] Luminosity L_{\star} ( \text{L}_{\odot} ) 0.265 \pm 0.011 § III.1 Interferometry Angular diameter \theta (mas) 1.035 \pm 0.021 R. Ligi et al. [112] Effective temperature T_{\text{eff}} (K) 4854 \pm 66 1 § III.1 Radius R_{\star} ( \text{R}_{\odot} ) 0.728 \pm 0.015 § III.1 Angular diameter \theta (mas) 1.114 \pm 0.007 A. Elliott et al. [57] Effective temperature T_{\text{eff}} (K) 4678 \pm 45 § III.1 Radius R_{\star} ( \text{R}_{\odot} ) 0.783 \pm 0.005 § III.1 Spectroscopy Effective temperature T_{\text{eff}} (K) 4817 \pm 62 L. J. Rosenthal et al. [158] Metallicity {\rm[Fe/H]} (dex) 0.08 \pm 0.06 1 L. J. Rosenthal et al. [158] Asteroseismology \Delta\nu ( \mu{\rm Hz} ) 182.799\pm 0.069 § II \epsilon 1.455\pm 0.010 § II \delta\nu_{0,1} ( \mu{\rm Hz} ) 3.81\pm 0.41 § II \delta\nu_{0,2} ( \mu{\rm Hz} ) 10.90\pm 0.41 § II \delta\nu_{0,3} ( \mu{\rm Hz} ) 22.27\pm 0.59 § II \nu_{\text{max}} ( \mu{\rm Hz} ) 4651\pm 301 § VII A_{\text{osc}} (cm s -1 ) 4.23\pm 0.41 § VII Mass M_{\star} ( \text{M}_{\odot} ) 0.763 \pm 0.020 (stat) \pm 0.014 (sys) § III.2 Radius R_{\star} ( \text{R}_{\odot} ) 0.748 \pm 0.007 (stat) \pm 0.004 (sys) § III.2 Age t_{\star} (Gyr) 10.2 \pm 1.5 (stat) \pm 1.0 (sys) § III.2 Density \rho_{\star} ( \rho_{\odot} ) 1.823 \pm 0.003 (stat) \pm 0.003 (sys) § III.2 Surface gravity {\log g}_{\star} (dex) 4.577 \pm 0.001 (stat) \pm 0.003 (sys) § III.2 Initial helium abundance Y_{\rm init} 0.27 \pm 0.02 (stat) \pm 0.01 (sys) § III.2 Rotation Rotation period P_{\rm rot} (d) 41.3 \pm 2.8 § V.1 HD 219134b Mass ( \text{M}_{\earth} ) 4.59 \pm 0.16 § VI Radius ( \text{R}_{\earth} ) 1.542 \pm 0.054 § VI HD 219134c Mass ( \text{M}_{\earth} ) 4.23 \pm 0.20 § VI Radius ( \text{R}_{\earth} ) 1.455 \pm 0.046 § VI Note. — 1. Used for asteroseismic modeling. III.2 Frequency modeling For frequency modeling, we employed five distinct modeling pipelines to assess the systematic uncertainties arising from different choices of input physics and the use of various codes. Stellar evolution codes used by the five teams included MESA [ 147 , 148 , 149 , 150 , 151 , 98 , 136 ] , GARSTEC [ 194 ] , and YREC [ 50 ] . Pulsation codes used for calculating oscillation frequencies included GYRE [ 183 ] , ADIPLS [ 39 ] , and H. M. Antia & S. Basu [5] . The input physics used by the five teams also differed. Treatments of atmospheric boundary conditions included the Eddington [ 56 ] and Krishna-Swamy [ 107 ] T - \tau relations, and pre-computed photosphere tables [ 81 , 82 , 31 , 185 ] . Choices of nuclear reaction rates varied from R. H. Cyburt et al. [46] , C. Angulo et al. [4] , A. Formicola et al. [63] , J. W. Hammer et al. [78] , and E. G. Adelberger et al. [1] . The mixing length formulations included L. Henyey et al. [83] , J. P. Cox & R. T. Giuli [42] and E. Böhm-Vitense [18] , with some teams using solar-calibrated mixing length parameters, others treating it as a free variable, and still others using calibrated prescriptions for varying it. Two main metal mixtures were used: those of M. Asplund et al. [6] and of N. Grevesse & A. J. Sauval [74] . Mass, metallicity, and initial helium abundance were all treated as free parameters by all teams. Different teams employed different correction procedures for near-surface modeling errors. Several of the teams used empirical correction formulae [ 12 ] , where an offset is applied to the mode frequencies to account for the effects of the surface term, before the corrected mode frequencies are used to construct a conventional likelihood function. These empirical formulae admit additional free parameters, which may either be fitted against the data, or potentially be externally calibrated against 3D models [ 173 ] or from an empirical ensemble [ 110 ] . Other teams employed the nonparametric \epsilon -matching technique [ 159 ] , where combinations of mode frequencies are computed from both the model and observed set, chosen so that the likelihood function evaluated from these combinations is insensitive to the near-surface structure of the star (see e.g. J. M. J. Ong et al. 146 for a review of surface-term correction techniques). We discuss in more depth in § IV.2 their potential systematic effects on our stellar-modeling results. Table 2 lists the detailed configurations as well as the derived stellar parameters by each team. Overall, the derived stellar properties for HD 219134 show strong consistency across modeling teams, despite intentional variations in input physics. The standard deviation of median values from each team is comparable to or smaller than the average uncertainties reported by each team, indicating that model uncertainties arising from differences in input physics and stellar evolution codes are generally smaller than the formal uncertainties reported. We report our stellar properties of HD 219134 as follows: we adopted the median values and the statistical uncertainties from Team 1’s median and formal uncertainty, and adopted the standard deviation of median values from each team as an estimate of the systematic uncertainties. The results are reported in Table 3 and Figure 4 . We find a mass of M_{\star}=0.763\pm 0.020({\rm stat})\pm 0.007({\rm sys}) \text{M}_{\odot} , a radius of R_{\star}=0.748\pm 0.007({\rm stat})\pm 0.002({\rm sys}) \text{R}_{\odot} , an age of t_{\star}=10.151\pm 1.520({\rm stat})\pm 0.810({\rm sys}) Gyr, and an initial helium abundance of {Y_{\rm init}}{}=0.27\pm 0.02({\rm stat})\pm 0.01({\rm sys}) . The fractional uncertainties of mass, radius, and age are 3%, 1%, and 18%, respectively. The derived Y_{\rm init} is close to the solar initial helium abundance ( \approx 0.27; M. Asplund et al. 6 ). Given that HD 219134’s metallicity is also close to solar, this Y_{\rm init} aligns well with expectations from Galactic enrichment laws [ 121 , e.g.,] . This constraint on Y_{\rm init} is derived largely indirectly from other constraints rather than directly from helium glitch signatures, as the latter typically produce signals below 1 \mu{\rm Hz} , which is smaller than the statistical uncertainties reported in Table 1 [ 190 , 126 ] . Furthermore, we determine a stellar density of \rho_{\star}=1.823\pm 0.003({\rm stat})\pm 0.003({\rm sys}) \rho_{\odot} , and a surface gravity of {\log g}_{\star}=4.577\pm 0.001({\rm stat})\pm 0.003({\rm sys}) dex. The systematic uncertainties are comparable to or larger than the statistical uncertainties, highlighting the need to account for both sources of error for these properties [ 94 , e.g.] . Figure 5: Stellar evolutionary models from zero-age to terminal-age main sequence, calculated by Team 1, presented on the C–D diagram (top; \delta\nu_{0,2} vs. \Delta\nu ) and the H–R diagram (bottom). The models are color-coded by stellar age. The position of HD 219134 is marked by a star symbol, with 1 \sigma uncertainties, some of which are smaller than the symbol size. III.3 Asteroseismic Age is Less Affected by Surface Physics Our estimate of the asteroseismic age ( t_{\star}=10.151\pm 1.520({\rm stat})\pm 0.810({\rm sys}) Gyr) is approximately 1- \sigma lower than the age derived by M. Gillon et al. [69] , who reported t_{\star}=11.0\pm 2.2 Gyr based on modeling that incorporated a precise interferometric radius. Unlike constraints based on radius, asteroseismic ages primarily depend on constraints on the star’s internal structure, particularly the hydrogen-burning processes near the core. Figure 5 presents stellar evolutionary models computed by Team 1 alongside the star’s position. On the H–R diagram, stellar tracks with different input parameters overlap at the star’s position, making it difficult to determine the stellar age based solely on surface properties without implementing any priors on the input parameters. However, the C–D diagram provides a clearer age diagnostic, as \delta\nu_{0,2} traces the fraction of hydrogen burned in the core [ 38 ] . Stellar properties derived using asteroseismic constraints is less influenced by assumptions in surface modeling or the accuracy of surface constraints such as radius and T_{\text{eff}} . For example, Teams 1 and 3 examined the effect of different choices of classical observables, either using T_{\text{eff}} alone or in combination with the interferometric radius from R. Ligi et al. [112] , along with asteroseismic constraints. They found that the resulting ages remained consistent within the quoted uncertainties. IV Radius Inflation In Table 3 and Figure 4 , we compare the asteroseismic radius derived in this study with the interferometric radius determined by A. Elliott et al. [57] . The asteroseismic radius derived for HD 219134, R_{\star}=0.748\pm 0.007({\rm stat})\pm 0.002({\rm sys}) \text{R}_{\odot} , is significantly lower than the new interferometric radius, R_{\star}=0.783\pm 0.005 \text{R}_{\odot} [ 57 ] . Both radii are highly precise, with uncertainties below 1%, yet they differ by approximately 4% (a 4 \sigma discrepancy). While the exact cause of this discrepancy is still unclear, in this section, we explore several explanations. Future observations of more targets common to both asteroseismology and interferometry could help identify the source of this discrepancy. IV.1 Systematic Uncertainties in Interferometry Interferometric visibilities can be affected by systematic errors due to assumed calibrator diameters, wavelength calibration, or limb-darkening corrections [ 187 , e.g.] . Systematic differences in interferometric angular diameters measured using different methods and instruments can reach up to 10% [ 198 , 179 ] . These systematic errors are most significant if the angular diameter of a target star is not well resolved compared to the calibrator diameters, which is supported by systematic trends between photometric and interferometric temperature scales [ 30 ] . R. Ligi et al. [112] used the optical VEGA beam combiner [ 141 ] on the CHARA Array [ 127 ] to measure the angular diameter of for HD 219134 using baselines up to \approx 100 m, yielding a minimum visibility of V^{2}\approx 0.2-0.3 . The latest interferometric data obtained using the optical PAVO beam combiner on the CHARA Array [ 95 ] used a longer baseline up to \approx 160 m, thus better resolving the star down to the first visibility null [ 91 , 57 ] . This makes the derived angular diameter essentially independent of systematics due to assumed calibrator sizes. Systematic effects due to limb darkening corrections and wavelength calibration are expected to be well below the 1% level [ 90 , 92 ] . Given previous good agreement between PAVO and VEGA for stars that are similarly well resolved [ 111 , 197 ] , we conclude that it is unlikely that the \approx 4% radius tension between interferometry and asteroseismology can be attributed to systematic errors in the interferometric measurements. Systematic errors in parallax scales by Gaia at this level can also be ruled out [ 75 ] . IV.2 Atmospheric Boundary Conditions We note that the majority of the asteroseismic constraints presented in this work are explicitly constructed to be insensitive to the surface (and therefore radius) of the star. This is because modeling errors — arising from the neglect of more realistic equilibrium structures in the stellar atmosphere, and the interaction between convection and oscillations — are known to produce a systematic offset between frequencies of actual stars, and those of notional 1D stellar models with identical internal structures. Corrections for this so-called “surface effect" must therefore be applied to model frequencies, in order to mitigate this discrepancy, but doing so often renders these asteroseismic constraints insensitive to the structure of the near-surface layers through the introduction of additional free parameters. Under these circumstances, the reported radius does not depend significantly on the precise form chosen to describe this surface effect using free parameters [ 144 , 146 ] . Conversely, this means that the radius estimates returned from these asteroseismic modeling efforts are primarily model-dependent, in that they are the radii of 1D models possessing interior structures consistent with the asteroseismic constraints, rather than radii directly constrained using seismology. All else being equal, the radii of 1D models in turn are determined by the atmospheric boundary conditions imposed on them computationally. We confirmed that the determined radius is not highly sensitive to some specific modeling choices. This was demonstrated by modifying only one aspect of the models while keeping all other physical assumptions fixed. To assess the influence of the atmospheric boundary condition, Team 1 replaced the photosphere tables [ 81 , 82 , 31 ] with the A. S. Eddington [56] model, resulting in a slight radius decrease of 0.006 \text{R}_{\odot} . Team 4 examined the effect of replacing A. S. Eddington [56] ’s T - \tau relation with that of K. S. Krishna Swamy [107] and found a small radius decrease of 0.009 \text{R}_{\odot} . Additionally, in implementing the surface correction procedure, Team 2 replaced their use of the calibrated T. Sonoi et al. [173] formulation (which has no free parameters) with that of W. H. Ball & L. Gizon [12] (which includes free parameters), leading to a minor radius reduction of 0.001 \text{R}_{\odot} . Moreover, the modeling teams employed a diverse set of atmospheric models, including T - \tau relations such as those of Eddington [ 56 ] and Krishna-Swamy [ 107 ] , and precomputed atmosphere tables based on more realistic 1D or 3D models [ 81 , 82 , 31 , 185 ] . Despite this broad range of approaches, the differences in the predicted radius across the modeling teams remained small, which suggests that the common choice of atmospheric boundary conditions and surface correction procedure fails to explain the radius discrepancy. IV.3 Mixing Lengths Figure 6: Probability contours of stellar properties presented in the space of radius, \alpha_{\rm MLT} and Y_{\rm init} , based on models calculated by Team 1. These contours are derived from constraints provided by asteroseismic observations. The radii of 1D stellar models with convective envelopes are determined by convective mixing-length theory, which specifies the temperature stratification of the superadiabatic layers close to the stellar surface. Since most of our modeling teams treated the mixing length parameter \alpha_{\rm MLT} either as a global free parameter or a fixed one according to calibrations against the Sun, incorrect values of \alpha_{\rm MLT} might have systematically biased the resulting asteroseismic model radius constraints. 3D hydrodynamical simulations produce entropy profiles associated with this temperature stratification that cannot be reproduced in 1D stellar evolution when \alpha_{\rm MLT} is treated as a globally constant quantity. Calibrated prescriptions for \alpha_{\rm MLT} also exist [ 174 , 185 , e.g.] , anchored to relations between the adiabatic entropy, surface gravity, and effective temperature that are seen to emerge in these simulations [ 123 , 176 ] . One modeling team (Team 5) performed their modeling using the variable- \alpha_{\rm MLT} prescription of [ 185 ] , but the radius and age scale of this modeling was still ultimately determined by a solar-calibrated value of {\alpha_{\rm MLT}}=1.83 , against which the local {\alpha_{\rm MLT}} was obtained by scaling under this prescription. There have also been suggestions that \alpha_{\rm MLT} should change as a function of stellar mass and/or metallicity in main-sequence dwarfs. 3D simulations indicate that the entropy jump between the bottom and top of the super-adiabatic region is smaller for cooler dwarfs [ 184 ] . This trend corresponds to higher convective efficiency and, consequently, a larger \alpha_{\rm MLT} [ 123 ] . These suggest that {\alpha_{\rm MLT}}{}/\alpha_{\text{MLT},\odot} at 4800 K is between 0.95 and 1.10 [ 185 , 123 ] , consistent with 2D simulations using gray radiative transfer [ 116 ] . 1D stellar models, when constrained by oscillation frequencies and atmospheric parameters, can also provide estimates of \alpha_{\rm MLT} [ 191 , e.g.,] . Observations of the well-characterized visual binary system \alpha Cen A and B further support the trend of increasing \alpha_{\rm MLT} with decreasing stellar mass [ 99 ] , even though the fit of \alpha_{\rm MLT} in 1D models allows the compensation for other deficiencies in the underlying model physics. These studies all support the notion that \alpha_{\rm MLT} at 4800 K should be near or slightly above solar. Accordingly, we investigated the shift of radius as a response to changing \alpha_{\rm MLT} . Figure 6 presents the posterior distributions of radius, \alpha_{\rm MLT} , and initial helium abundance ( Y_{\rm init} ) using Team 1’s stellar models. We observe a correlation between \alpha_{\rm MLT} and radius from the distributions. However, allowing a 20% change in \alpha_{\rm MLT} from the best-fitting value still fails to reconcile the predicted radius with that derived from interferometry. Furthermore, larger values of \alpha_{\rm MLT} reduce Y_{\rm init} to values near or below the primordial helium abundance of 0.249 [ 154 ] , which is physically implausible. Furthermore, the use of different input physics (investigated by this paper) is unlikely to shift \alpha_{\rm MLT} significantly. We see near-solar \alpha_{\rm MLT} across modeling teams where \alpha_{\rm MLT} was treated as a free parameter: averaging around {\alpha_{\rm MLT}}/{\alpha_{\rm MLT}}_{\odot}=0.980\pm 0.053 . We also found that the \alpha_{\rm MLT} values are not significantly affected by changes in their boundary conditions. This is supported by comparisons between models that differ only in their boundary conditions: Team 1 repeated their analysis, replacing their photosphere tables with the A. S. Eddington [56] boundary condition, and Team 4 replaced the A. S. Eddington [56] boundary condition with that of K. S. Krishna Swamy [107] . Both teams found the changes in {\alpha_{\rm MLT}}{}/{\alpha_{\rm MLT}}{}_{\odot} remained within their quoted 1 \sigma confidence intervals. IV.4 Magnetic Fields K and M dwarfs are often observed to have larger radii than those predicted by stellar models, a phenomenon referred to as radius inflation [ 182 ] . This discrepancy can be resolved in models by incorporating magnetic fields directly in stellar structure equations [ 58 , 122 , 59 ] , or by including starspots that inhibit energy transport on the stellar surface [ 171 ] . Observational evidence suggests that the degree of radius inflation correlates with surface magnetic field strength [ 101 ] and rotation period [ 108 ] , supporting the role of magnetic activity as a key factor. Modeling studies found that strong magnetic fields are necessary to account for radius inflation. G. Somers & M. H. Pinsonneault [171] found that a 50% spot coverage is required to explain a 4% radius difference (see their Figure 2). G. A. Feiden & B. Chaboyer [59] studied UV Psc B, which shows a 10% radius inflation compared to non-magnetic models, requiring surface magnetic fields of approximately 4 kG to match the observations. However, HD 219134 appears not to possess a magnetic field strong enough to explain the observed discrepancy. The surface magnetic field strength measured by C. P. Folsom et al. [61] is on the order of 2.5 G, comparable to the Sun and the \alpha Cen A and B systems [ 58 ] . Furthermore, observations of the star’s chromospheric emission using Ca II H and K lines spanning over a decade (see § V.4 ) places HD 219134 firmly in the inactive regime of stellar activity distribution. Thus, magnetic activity is very unlikely to account for the 4% difference between the asteroseismic and interferometric radii. IV.5 Tidal Heating The orbital angular momentum of a planet can drive tidal interactions within the convective envelope of its host star, leading to energy dissipation in the form of heat. This dissipation can act as an additional energy source, potentially contributing to the inflation of the host star. However, in the case of HD 219134, which hosts two ultra-short-period super-Earths, HD 219134b and HD 219134c, the expected tidal effect is insufficient to account for a 4% increase in the stellar radius. To examine this, we calculated the tidal decay timescale, \tau_{d} , for a slowly rotating host star using the following expression [ 199 , 72 , 48 ] : \tau_{d}\approx 30~\text{Gyr}\left(\frac{Q^{\prime}_{\star}}{10^{6}}\right)\left(\frac{M_{\star}/{\text{M}_{\odot}}}{M_{p}/{\text{M}_{\earth}}}\right)\left(\frac{\rho_{\star}}{\rho_{\odot}}\right)^{5/3}\left(\frac{P_{\rm orb}}{1~\text{day}}\right)^{13/3}, (5) where Q^{\prime}_{\star} is the tidal quality factor, which we assume to be 10^{7} [ 152 ] , and P_{\rm orb} is the orbital period. Using this timescale, we derived the corresponding energy dissipation rate: \dot{E}=\frac{3}{2}\frac{GM_{\star}M_{p}}{P_{\rm orb}}\frac{1}{\tau_{d}}. (6) With the stellar and planetary parameters from Table 3 and M. Gillon et al. [69] , we determined that the energy dissipation induced by the close-in planet HD 219134b is 1.17\times 10^{-7}~{\text{L}_{\odot}}{} . Even if all this dissipated energy were converted into stellar luminosity, it would still fall several orders of magnitude short of explaining the required 8% luminosity discrepancy (corresponding to the 4% difference in radius). V Evolution of Rotation and Activity Using the estimated asteroseismic age, t_{\star}=10.2\pm 1.5({\rm stat})\pm 1.0({\rm sys}) Gyr, alongside an independently measured rotation period, we can test the theory of angular momentum loss. Figure 7: Time series of \log R^{\prime}_{\rm HK} measured by three different instruments spanning 20 years (HARPS, F. Motalebi et al. 140 ; HIRES and APF, L. J. Rosenthal et al. 158 , H. Isaacson et al. 96 ). Figure 8: Calibrators of gyrochrones shown on the P_{\rm rot} – T_{\text{eff}} diagram. The open cluster data include Pleiades [ 155 ] , Praesepe [ 52 , 53 ] , NGC 6811 [ 44 ] , NGC 752 [ 3 ] , NGC 6819 [ 129 ] , Ruprecht 147 [ 45 ] , and M67 [ 13 , 55 , 114 ] . The asteroseismic field stars shown include stars with both rotation periods and ages measured via asteroseismology [ 77 , 169 , 170 ] . The background displays a binned distribution of the Kepler field star sample [ 128 ] . HD 219134 is the coolest age calibrator for stars older than 4.2 Gyr. Figure 9: Rotation evolutionary models projected to older ages. These are calibrated based on asteroseismic field stars and open clusters hotter than 4500 K. The calibrators within the {T_{\text{eff}}}\in[4650,4900] K range are shown in triangles. The evolutionary tracks include models within 3- \sigma that match HD 219134’s mass and metallicity. Additional model configurations are specified in each panel. Models highlighted in dark gray represent those that have entered the weakened magnetic braking phase ( \differential J_{\rm tot}/\differential t=0 ). V.1 Rotation Period Based on Calcium H and K line emission measured with HARPS-N over three years, F. Motalebi et al. [140] reported a rotation period for HD 219134 of {P_{\rm rot}}=43.2 d. Similarly, C. P. Folsom et al. [61] , using Zeeman–Doppler imaging from Narval spectropolarimeter, found a rotation period of {P_{\rm rot}}=42 d. The spread in rotation period estimates from different methods is consistent with expected effects of latitudinal differential rotation. We used HIRES and APF data collected via California Planet Search [ 158 , 96 ] , as well as HARPS data [ 140 ] to determine the star’s rotation period. These data are shown in Figure 7 . We modeled the time series using a hierarchical model and Gaussian processes [ 62 ] . Each time series was assigned a single rotation period, with all periods assumed to follow a normal distribution with a common mean. As a result, we determined a rotation period of {P_{\rm rot}}=41.3\pm 2.8 d. Figure 8 shows the position of this star on the P_{\rm rot} – T_{\text{eff}} plane, compared with other stars of known rotation period and benchmark ages derived from open clusters or asteroseismology. Stellar angular momentum loss over time is driven primarily by magnetized stellar winds. Establishing a well-calibrated rotation-age relationship is essential for age-dating cool dwarf stars, especially for exoplanet hosts. HD 219134 is noteworthy for occupying a unique parameter space as the coolest benchmark star with an age exceeding 4.2 Gyr. V.2 Rotation Modeling To check whether our current knowledge of magnetic braking and angular momentum transport is sufficient to explain the rotation period of HD 219134 at 10.151 Gyr, we used the stellar evolution models calculated from Team 1, and calculated stellar rotation using ROTEVOL [ 189 , 172 ] . Models were initialized with a rotation period of P_{\rm disk}=4 d and disk-locking timescale of \tau_{\rm disk}=10 Myr, following the choices of F. Chiti et al. [35] . We employed a magnetic braking model based on J. L. van Saders et al. [188] , which quantifies the total angular momentum loss as: \derivative{J_{\rm tot}}{t}=\left\{\begin{aligned} &f_{K}K_{M}\omega_{e}\left(\frac{\omega_{\rm sat}}{\omega_{\odot}}\right)^{2},\ \omega_{\rm sat}\leq\omega_{e}\frac{\tau_{\rm cz}}{\tau_{\rm cz,\odot}},{{\rm Ro}}{}\leq{{\rm Ro}}_{\rm crit}\\ &f_{K}K_{M}\omega_{e}\left(\frac{\omega_{e}\tau_{\rm cz}}{\omega_{\odot}\tau_{\rm cz,\odot}}\right)^{2},\ \omega_{\rm sat}>\omega_{e}\frac{\tau_{\rm cz}}{\tau_{\rm cz,\odot}},{{\rm Ro}}{}\leq{{\rm Ro}}_{\rm crit}\\ &0,\ {{\rm Ro}}>{{\rm Ro}}_{\rm crit}\end{aligned}\right. (7) Here, \omega_{e} is the envelope rotation rate, \tau_{\rm cz} is the convective overturn timescale, calculated locally at one pressure scale height above the bottom of the outer convection zone, \omega_{\rm sat}=3.863\times 10^{-5} rad s -1 is the saturation threshold of angular momentum loss, {\rm Ro} is the Rossby number defined as {P_{\rm rot}}/\tau_{\rm cz} as a means to estimate stellar magnetism, and {{\rm Ro}}_{\rm crit} is a Rossby threshold, beyond which no angular momentum loss occurs [ 188 , 163 ] . The factor K_{M} scales with stellar properties as: \frac{K_{M}}{K_{M,\odot}}=c\left(\frac{R}{R_{\odot}}\right)^{3.1}\\ \left(\frac{M}{M_{\odot}}\right)^{-0.22}\left(\frac{L}{L_{\odot}}\right)^{0.56}\left(\frac{P_{\rm phot}}{P_{\rm phot,\odot}}\right)^{0.44}, (8) where c corrects for centrifugal force (assumed to be 1), and P_{\rm phot} denotes the photospheric pressure. We modeled the stellar rotation using a two-zone model, with the core and envelope rotating independently as solid bodies, and exchanging angular momentum over a coupling timescale \tau_{\rm ce} . This timescale scales with stellar mass as \tau_{\rm ce}/\tau_{\rm ce,\odot}=(M/M_{\odot})^{-\alpha_{\rm ce}} , with \tau_{\rm ce,\odot}\approx 22 Myr [ 175 ] . The angular momentum loss rates for the core and envelope are expressed as [ 51 ] \derivative{J_c}{t}=-\frac{(\Delta J)_{\rm max}}{\tau_{c}}-\omega_{s}\derivative{I_e}{t}, (9) \derivative{J_e}{t}=\frac{(\Delta J)_{\rm max}}{\tau_{c}}+\omega_{s}\derivative{I_e}{t}+\derivative{J_{\rm tot}}{t}, (10) where I_{e} and I_{c} represent envelope and core inertia, respectively. The term \omega_{s} is defined as the envelope rotation rate \omega_{e} if the envelope is contracting, or the core rotation rate \omega_{c} if the envelope is expanding. The maximum angular momentum exchange between the core and envelope is given by: (\Delta J)_{\rm max}=\frac{I_{c}I_{e}}{I_{c}+I_{e}}(\omega_{c}-\omega_{e}). (11) The free parameters in this model are f_{K} , {{\rm Ro}}_{\rm crit}/{{\rm Ro}}_{\odot} and \alpha_{\rm ce} . To calibrate these parameters, we constructed a sample of open clusters, field stars, and the Sun with their rotation periods and ages. The open clusters were the Pleiades [ 155 , 120 Myr;] , Praesepe [ 52 , 53 , 670 Myr;] , NGC 6811 [ 44 , 1 Gyr;] , NGC 6819 [ 129 , 2.5 Gyr;] , Ruprecht 147 [ 45 , 2.7 Gyr;] , and M67 [ 13 , 55 , 114 , 4.2 Gyr;] . The field stars were Kepler stars with rotation periods [ 77 ] and asteroseismic ages [ 169 , 170 ] . Finally, the Sun was included as a data point with {P_{\rm rot}}{}=25.4 d and t_{\odot}=4.67 Gyr [ 9 ] . We restricted the sample to stars with effective temperatures above 4500 K, to avoid difficulties in modeling stars around 4200 K, which have enhanced activity that is thought to be due to shear-enhanced magnetism [ 27 , e.g.] . The best-fitting model parameters obtained are f_{K}=5.655 , {{\rm Ro}}_{\rm crit}/{{\rm Ro}}_{\odot}=0.93 , and \alpha_{\rm ce}=3.8 . V.3 Testing Rotation Models Beyond 4.2 Gyr We searched these rotational evolutionary tracks for models with the same mass and metallicity as HD 219134 by predicting rotation periods to the age of HD 219134. Figure 9 (a) presents these evolutionary tracks, calculated using the default weakened magnetic braking configuration. These models assume initial helium abundance ( Y_{\rm init} ) and mixing length parameter ( \alpha_{\rm MLT} ) constrained by asteroseismology, along with the best-fitting core-envelope coupling coefficient ( \alpha_{\rm ce} ). The models agree very well with HD 219134’s observed rotation period and age. The estimated {{\rm Ro}}/{{\rm Ro}}_{\odot} is 0.83 \pm 0.07, only slightly below the critical value {{\rm Ro}}_{\rm crit}/{{\rm Ro}}_{\odot}=0.93 , indicating the star is around the transition point into the weakened magnetic braking regime. H. Isaacson et al. [96] reported an activity cycle of {P_{\rm cyc}}=13.27 yr for HD 219134, placing it on the lower sequence in the P_{\rm cyc} – P_{\rm rot} diagram. Following T. S. Metcalfe et al. [133] , this positioning also implies HD 219134 is close to the transition into weakened magnetic braking [ 188 ] , consistent with our findings. In fact, T. S. Metcalfe et al. [134] found evidence of significantly weakened instantaneous torque in HD 219134, based on constraints on magnetic field strength and morphology from ZDI, suggesting the star has already entered the weakened magnetic braking phase. Given that our estimate of {{\rm Ro}}/{{\rm Ro}}_{\odot} is 1\sigma below {{\rm Ro}}_{\rm crit}/{{\rm Ro}}_{\odot} , this provides tentative (though not statistically significant) evidence that {{\rm Ro}}_{\rm crit} for this star may be different from most G-dwarf asteroseismic calibrators. We also modeled rotational evolution assuming standard magnetic braking, which prescribes continued spin-down even when {{\rm Ro}}>{{\rm Ro}}_{\rm crit} (modifying the third line in Equation 7 to equal the second line). The results are shown in Figure 9 (b). As expected, this model predicts further spin-down beyond 10 Gyr. If the observation of instantaneous torque is not considered, both the standard and weakened braking models adequately explain the current observations. This is because the divergence in rotation periods between the two regimes requires a longer timescale to become clearly distinguishable. Next, we tested whether models with solar-calibrated Y_{\rm init} and \alpha_{\rm MLT} could also explain the observations. These results are shown in Figure 9 (c). Compared with Figure 9 (a), models with Y_{\rm init} and \alpha_{\rm MLT} constrained by asteroseismology provide a better overall fit. The initial helium abundance ( Y_{\rm init} ) affects the stellar structure by altering the size of the convection zone, and hence \tau_{\rm cz} [ 163 ] , similar to the effects of [Fe/H] or [ \alpha /M] [ 41 ] . This highlights the importance of incorporating detailed chemical abundances for accurately predicting ages using gyrochronology. Finally, we explored models with a large coupling timescale motivated by studies such as L. Cao et al. [27] , who found \alpha_{\rm ce}\approx 12 for late K dwarfs with enhanced magnetism in the Praesepe open cluster. These models are presented in Figure 9 (d). While the rotation period and age of HD 219134 agree with the model, the rotation period for stars around 3 Gyr is underestimated. Thus, our data do not support a larger coupling timescale for stars at this T_{\text{eff}} . This suggests that the scaling relation for coupling timescale versus mass may be oversimplified for modeling the stalled spin-down phase. In summary, weakened magnetic braking models with constrained abundances provide a good match to HD 219134’s observed rotation period and asteroseismic age. Our findings support the reliability of such models for predicting rotational evolution in early K-type dwarfs beyond 4.2 Gyr. The effects of prolonged core-envelope coupling earlier in stellar evolution can leave observable imprints in older stars. Combining asteroseismic ages and torque measurements for a statistically significant sample will help constrain the coupling timescale as a function of T_{\text{eff}} . V.4 Activity-Age Relations Figure 10: Chromospheric activity-age relation. The sample includes solar twins [ 115 ] , the Kepler asteroseismic sample [ 100 , 130 , 43 ] , and the TESS asteroseismic sample [ 94 , 37 , 131 , 132 ] . A common method for age-dating field stars involves using activity-age relations. These activity indices include metrics such as spot modulation amplitudes, emissions in spectral lines (e.g., Calcium H&K and H \alpha ), and emissions in the X-ray or radio spectrum. In this study, we examined whether the chromospheric activity of HD 219134 aligns with established activity-age relationships. Wherever possible, we used the color-temperature relation from M. Sekiguchi & M. Fukugita [167] and converted the S-indices into the color-independent activity index \log R^{\prime}_{\rm HK} , on the scale defined by F. Middelkoop [135] and R. W. Noyes et al. [143] . Figure 10 presents \log R^{\prime}_{\rm HK} for a sample of solar twins [ 115 ] and asteroseismic samples from Kepler and TESS [ 100 , 130 , 43 , 94 , 37 , 131 , 132 ] . We found that at an age of 10 Gyr, HD 219134 exhibits a higher activity level compared to solar twins of the same age but is comparable to the activity levels observed in solar twins at 5 Gyr. This is consistent with expectations that activity levels scale with {\rm Ro} , as {\rm Ro} of HD 219134 is close to {{\rm Ro}}_{\rm crit} , a value also observed for solar twins at 5 Gyr, beyond which they enter into the phase of weakened magnetic braking. This result supports the inclusion of a T_{\text{eff}} dependence in the \log R^{\prime}_{\rm HK} –age relation, particularly for cooler stars, as differences in {\rm Ro} can affect the activity-age trend by as much as 5 Gyr. VI Revising Planet Properties Figure 11: The mass–radius diagram of exoplanets with R_{p}<1.8~{\text{R}_{\earth}} , using planet properties from the NASA Exoplanet Archive [ 142 ] . Ultra-short-period planets are defined following F. Dai et al. [47] , and are expected to be bare rocky cores whose composition can be probed directly with mass and radius measurements. The solid lines indicate different theoretical compositional models calculated by L. Zeng et al. [202] . HD 219134 hosts five confirmed exoplanets, including two transiting super-Earths (HD 219134b and HD 219134c) first discovered by M. Gillon et al. [69] from photometric analysis using the Spitzer Space Telescope. As one of brightest star known to host transiting planets, the system has been extensively studied, including the possibility of the formation and detection of planetary exospheres [ 192 ] . We produce revised properties of the two transiting planets based on our new asteroseismic stellar mass and radius. The mass of transiting planets can be calculated using M_{p}=\frac{K_{1}^{2}+\sqrt{K_{1}^{4}+4G(1-e^{2})^{-1}\sin^{2}i\ a^{-1}K_{1}^{2}M_{\star}}}{2G(1-e^{2})^{-1}\sin^{2}i\ a^{-1}}, (12) where K_{1} , e , i , a , G denote the RV semi-amplitude, orbital eccentricity, orbital inclination, semi-major axis, and the gravitational constant, respectively. The radius of a transiting planet is given by R_{p}=R_{\star}\cdot R_{p}^{\prime}/R_{\star}^{\prime}, (13) where R_{p}^{\prime}/R_{\star}^{\prime} is the old planet and star radius ratio determined by M. Gillon et al. [69] . Using the orbital parameters provided in M. Gillon et al. [69, see their Table 1] along with the asteroseismic mass and radius derived in this work, we revise the HD 219134b’s mass and radius from M_{p}=4.73\pm 0.17 \text{M}_{\earth} and R_{p}=1.605\pm 0.055 \text{R}_{\earth} to M_{p}=4.59\pm 0.16 \text{M}_{\earth} and R_{p}=1.542\pm 0.054 \text{R}_{\earth} . Similarly, we revise the mass and radius of HD 219134c from M_{p}=4.36\pm 0.21 \text{M}_{\earth} and R_{p}=1.512\pm 0.046 \text{R}_{\earth} to M_{p}=4.23\pm 0.20 \text{M}_{\earth} and R_{p}=1.455\pm 0.046 \text{R}_{\earth} . We plot both the previous and updated planet properties on the mass-radius diagram in Figure 11 . The revised mass and radius shift the original data points slightly downward, supporting the hypothesis that ultra-short-period super-Earths typically exhibit an Earth-like composition of 30% Fe and 70% MgSiO 3 composition [ 54 , 47 , 22 ] . VII Oscillation Amplitudes Figure 12: Oscillation amplitudes in radial velocity as a function of L/M (top) and T_{\text{eff}} (bottom) for main-sequence dwarfs, color-coded by chromospheric activity. Using pySYD , we measured the mode amplitude A_{\text{osc}} and the frequency of maximum power \nu_{\text{max}} from a heavily smoothed power spectrum. We obtained {A_{\text{osc}}}=4.23\pm 0.41 cm/s, and {\nu_{\text{max}}}=4651\pm 301 \mu{\rm Hz} . The power spectrum of HD 219134 appears to show a dip in the center of the oscillation region that is remarkably similar to that seen in \alpha Cen B, which the only other K dwarf with a high-SNR power spectrum (see Figure 4 of H. Kjeldsen et al. 105 ). Similar features are also seen in \epsilon Ind [ 26 , Figure 3 of] , and \sigma Dra [ 87 , Figure 6 of] , though based on shorter time series. It remains to be seen whether this is a real feature of K dwarfs, which would be important for our understanding of the excitation and damping mechanisms of solar-like oscillations [ 89 , see review by] . Figure 12 places HD 219134 among a sample of cool dwarf stars with oscillation amplitudes measured from radial velocity. The pulsation amplitude in radial velocity is theoretically expected to scale as {A_{\text{osc}}}\propto(L/M)^{s} , with s ranging from 0.7 to 1.5 [ 40 , 102 , 88 , 162 ] . Figure 12 also overlays MIST isochrones [ 36 ] to illustrate the relationship between A_{\text{osc}} and T_{\text{eff}} , assuming A_{\text{osc}} scales with (L/M) at power indices of 0.7 and 1.5, evaluated at ages of 1 and 4.6 Gyr, respectively. While an s index of 0.7 is consistent with observations of G- and F-type stars, T. L. Campante et al. [26] first suggested that cooler stars, such as \epsilon Indi, align more closely with s\approx 1.5. Our analysis of HD 219134 supports this result. It has also been proposed that higher stellar activity can reduce oscillation amplitudes. However, despite HD 219134 and \tau Ceti being relatively inactive, they follow the s\approx 1.5 relation, indicating that activity alone is unlikely to account for the steeper amplitude scaling observed in cooler stars. VIII Discussion and Conclusion In this work, we present an asteroseismic analysis of HD 219134 using 4 consecutive nights of radial velocity data obtained with Keck Planet Finder on Keck-I. We used the Gold deconvolution algorithm to disentangle the power spectrum from the spectral window, in order to permit identification of modes separated by close to 1\ \mathrm{c/d} , coinciding with the characteristic spacing of sidelobes. A total of 25 oscillation modes with spherical degrees 0\leq\ell\leq 3 were identified and extracted, as summarized in Table 1 . We then estimated HD 219134’s properties using five independent asteroseismic modeling pipelines. The resulting parameters are: mass 0.763 \pm 0.020 (stat) \pm 0.007 (sys) \text{M}_{\odot} , radius 0.748 \pm 0.007 (stat) \pm 0.002 (sys) \text{R}_{\odot} , and age 10.151 \pm 1.520 (stat) \pm 0.810 (sys) Gyr. These in turn led to revised masses and radii of the known transiting planets orbiting HD 219134, as detailed in Table 3 . Asteroseismology with EPRV instruments is a promising avenue for calibrating gyrochronology and studying angular momentum loss in old K and M dwarfs. HD 219134 in particular is the first dwarf cooler than 5000 K with an asteroseismic age estimate available. We used this to test models of rotational evolution used for this purpose by confronting them, at this age, with an independent measurement of its present rotational period. We demonstrated that existing models of angular momentum loss, incorporating weakened magnetic braking and asteroseismically constrained values of Y_{\rm init} and \alpha_{\rm MLT} , accurately reproduce HD 219134’s rotation period at its asteroseismic age. We further found that the spin-down relation for stars at this temperature requires a shorter spin-down coupling timescale than those suggested for late K dwarfs. Nonetheless, we note that this agreement, and ultimately the apparent robustness of our asteroseismic age estimate, requires that our evolutionary stellar modeling be accurate. Such an assumption of accuracy in modeling also fundamentally underpins our asteroseismic radius constraints, since these are (as we discuss in § IV.2 ) the radii of stellar models consistent with asteroseismic constraints on stellar interiors, rather than stellar radii directly constrained using asteroseismology. Importantly, we have also found our asteroseismic radius to be 4% smaller than the interferometric radius at 4- \sigma level, possibly indicating some deficiency in the accuracy of this modeling. We were unable to easily attribute this discrepancy to any systematic uncertainties related to interferometry, nor to variations in canonical choices of atmospheric boundary conditions or mixing length theory used in stellar modeling, nor to magnetic fields, nor tidal heating. Without any insight into the cause of this discrepancy, our subsequently derived quantities, and treatment of rotational evolution — all of which are contingent on these model ages and radii — must necessarily be regarded as being only conditional, pending a better understanding of the physical origin for this discrepancy. Future direct constraints on stellar radii from asteroseismology (e.g. through potential breakthroughs in understanding and mitigating the surface term) may alleviate this dependence on evolutionary modeling. Finally, we have confirmed that HD 219134’s oscillation mode amplitudes scale with those of other cool dwarfs as (L/M)^{1.5} , deviating from the (L/M)^{0.7} relation observed in G-type stars. Should this difference in scaling behaviour be genuine and general, this may have important implications for the interpretation of other asteroseismic observables, namely {\nu_{\text{max}}} , in similar stars. In particular, both the oscillation amplitudes and {\nu_{\text{max}}} are thought to be fully determined by the properties of turbulent stresses in the near-surface layers — both ultimately reflect the physical processes that excite normal modes to visible amplitudes in the first place. If the convective turbulence that operates G-type stars should indeed be fundamentally different in some fashion from in cooler dwarfs — yielding oscillation amplitudes that scale differently in these two classes of stars — then we might also expect a different {\nu_{\text{max}}} scaling relation to be required for these cooler dwarfs. Suggestively, HD 219134 joins \alpha Cen B [ 105 ] , \epsilon Ind [ 26 ] , and \sigma Dra [ 87 ] , in appearing to exhibit a broad excitation envelope that has a local minimum in mode amplitudes close to {\nu_{\text{max}}} , differing qualitatively from the solar phenomenology of a single, largely symmetric, envelope. Further, more detailed analyses of non-adiabatic pulsations and convective excitation, and a larger observational sample of likewise cool dwarfs, will be needed both to determine the exact cause of this difference, as well as to determine whether the solar-calibrated {\nu_{\text{max}}} scaling relation indeed also holds for stars with modified amplitude scaling. Appendix A Radial Velocity Data In Table 4 , we list the full radial velocity data measured from the KPF spectra obtained in this work. The raw KPF spectra can be accessed from Keck Observatory Archive. Table 4: Radial velocity data of HD 219134 measured from KPF spectra. BJD -2460579.72 (d) RV (m \cdot s -1 ) RV error (m \cdot s -1 ) 0.00946579 -18372.4687 0.4382 0.01016011 -18371.5709 0.4500 0.01085437 -18371.2015 0.4398 0.01154667 -18372.0286 0.4438 0.01222188 -18371.9946 0.4491 0.01292862 -18371.1237 0.4518 0.01360698 -18371.4315 0.4257 0.01429799 -18371.9583 0.4265 0.01500247 -18371.9011 0.4265 0.01567058 -18371.1193 0.4272 … … … Note. — Only the first 10 lines are shown. The full table can be accessed online.