DYNAMO-VLBA: Dynamical Masses of Young Stars

• DYNAMO-VLBA Project is an observational campaign that uses VLBA microarcsecond astrometry to provide model-independent dynamical mass measurements for tight pre-main sequence multiple systems.
• It employs multi-epoch VLBI observations to resolve orbits in binaries with separations of 1–30 mas, enabling empirical calibration of 0.2–6 M☉ pre-main sequence evolutionary tracks.
• Case studies of systems like S1 in Ophiuchus and EC 95 in Serpens reveal significant discrepancies with theoretical models, highlighting the need to recalibrate PMS evolutionary tracks.

The DYNAMO-VLBA project (“Dynamical Masses of Young Stellar Multiple Systems with the VLBA”) delivers high-precision, model-independent dynamical mass determinations for tight pre-main sequence (PMS) multiple systems in nearby star-forming regions, using microarcsecond (μas) astrometry obtained with the Very Long Baseline Array (VLBA). Targeting systems with component separations of 1–30 mas that are unresolved by Gaia, the project employs multi-epoch VLBI observations to resolve astrophysical orbits and benchmark pre-main sequence evolutionary models in the 0.2–6 $M_\odot$ regime. DYNAMO-VLBA has produced the most precise dynamical masses to date for the Ophiuchus Herbig Be binary S1 and the Serpens hierarchical triple EC 95, revealing critical validation and tension points for contemporary PMS tracks.

1. Scientific Motivation and Scope

The formation and early evolution of stars depend sensitively on mass—a fundamental stellar parameter. For young PMS stars (with ages $\lesssim$ 1 Myr), masses are generally inferred using theoretical isochrones in the Hertzsprung–Russell diagram (HRD). However, empirical calibration of these tracks, especially in the intermediate-mass range ($2{-}8\,M_\odot$), remains sparse. Multiple, tight binaries and higher-order systems are particularly valuable as they allow for direct dynamical mass measurement through orbital motion. Gaia astrometry saturates at separations $\lesssim$50 mas, precluding individual orbits for most such systems. By exploiting the unique angular resolution (down to $\sim$1 mas) and $\mu$as-level astrometric precision of the VLBA at $\nu=5$ GHz, DYNAMO-VLBA provides:

• Model-independent trigonometric distances and proper motions for young PMS multiples.
• Direct measurement of total and individual component masses from full orbital solutions.
• Empirical anchoring and potential recalibration of 0.2–6 $M_\odot$ PMS evolutionary tracks.

2. Target Selection and Observational Design

DYNAMO-VLBA selects systems exhibiting non-thermal radio emission (fluxes $\gtrsim$ 0.2 mJy at 5 GHz), typical of strong magnetic activity in young stellar objects. The survey focuses on nearby ($d=100{-}500$ pc), high-density star-forming regions (notably Ophiuchus and Serpens), where prior photometric or spectroscopic indicators of binarity suggest projected separations of 1–30 mas ($\lesssim$00.1–5 AU). Each selected system undergoes multi-epoch VLBA continuum monitoring:

• Array configuration: 10 antennas, $\lesssim$1 GHz dual polarization, 4 hr on-source per epoch.
• Phase-referencing cycle: 3 min science target / 1 min calibrator, for temporal atmospheric and clock error correction.
• Achieved resolution: $\lesssim$21.5 × 0.5 mas synthesized beam; image rms floor 10–20 $\lesssim$3Jy beam$\lesssim$4.
• Temporal sampling: typically 30–35 epochs per system across $\lesssim$5 10 yrs, ensuring full orbital phase coverage for binaries (e.g., S1 in Ophiuchus: 35 epochs, 2005–2017; EC 95 in Serpens: 32 epochs, 2007–2019).

For triple systems (EC 95), select epochs are supplemented with archival infrared astrometry to constrain tertiary component orbits.

3. Calibration and Astrometric Methodology

Data calibration employs the AIPS software suite, using standardized VLBI recipes:

• Fringe fitting on bright, nearby calibrators to derive delay and rate solutions.
• Amplitude and bandpass calibration against strong continuum calibrators.
• Transfer of calibration to target scans, phase-referenced with a cycle designed to minimize tropospheric residuals.

Source positions are determined through two-dimensional Gaussian fitting (AIPS task JMFIT), yielding right ascension and declination centroids and their associated uncertainties. Relative positional errors between binary components are calculated by quadrature summation of individual errors, resulting in typical absolute position uncertainties of 30–60 $\lesssim$6as.

Astrometric and orbital fitting is performed using simultaneous least-squares minimization (MPFIT), modeling each component’s trajectory as: $\lesssim$7 where $\lesssim$8 are the parallax factors, $\lesssim$9 represent the orbital displacements in Thiele-Innes formalism, and $2{-}8\,M_\odot$0 is the semi-major axis for each component. The model simultaneously fits the trigonometric parallax $2{-}8\,M_\odot$1, proper motion components $2{-}8\,M_\odot$2, and the full set of Keplerian elements (period, eccentricity, inclination, node, argument of periastron, time of periastron).

Dynamical masses are derived by: $2{-}8\,M_\odot$3 with angular-to-physical unit conversion $2{-}8\,M_\odot$4. Propagated errors follow

$2{-}8\,M_\odot$5

Secondary approaches, such as Orbitize! MCMC (Blunt et al. 2020), are used as independent consistency checks and for hierarchical systems (e.g., EC 95C).

4. Principal Results: Case Studies of S1 (Ophiuchus) and EC 95 (Serpens)

S1 (Ophiuchus)

• Parallax: $2{-}8\,M_\odot$6 mas $2{-}8\,M_\odot$7 pc.
• Orbit: $2{-}8\,M_\odot$8 mas ($2{-}8\,M_\odot$9 AU), $\lesssim$0 yr, $\lesssim$1, $\lesssim$2.
• Masses: S1A ($\lesssim$3 $\lesssim$4), S1B ($\lesssim$5 $\lesssim$6).
• Spectral Energy Distribution (SED): S1A is consistent with $\lesssim$7 K and $\lesssim$8. Luminosity from SED fitting is $\lesssim$9 $\sim$0.
• Notable result: S1A's measured mass is $\sim$125% less than the $\sim$26 $\sim$3 predicted by non-rotating and rotating PMS tracks at its HRD location; this discrepancy remains even after accounting for rotation effects. S1B is consistent with a low-mass ($\sim$40.8 $\sim$5), young T Tauri star.

EC 95 (Serpens)

• Parallax: $\sim$6 mas $\sim$7 pc.
• Inner binary (A–B): $\sim$8 mas ($\sim$9 AU), $\mu$0 yr, $\mu$1, $\mu$2.
• Masses: EC 95A ($\mu$3 $\mu$4), EC 95B ($\mu$5 $\mu$6).
• Tertiary (C): Orbit constrained by 4 IR + VLBA points, $\mu$7 yr, $\mu$8, $\mu$9; mass estimate $\nu=5$0 $\nu=5$1.
• Architecture: hierarchical triple; inner and outer orbit inclinations are similar ($\nu=5$235°), both orbits are highly eccentric.

A side-by-side summary of parameters is presented below:

System	S1A Mass ($\nu=5$3)	S1B Mass ($\nu=5$4)	EC 95A Mass ($\nu=5$5)	EC 95B Mass ($\nu=5$6)	EC 95C Mass ($\nu=5$7)
S1	4.11 ± 0.10	0.831 ± 0.014	–	–	–
EC 95	–	–	2.15 ± 0.10	2.00 ± 0.12	0.26$\nu=5$8

5. Implications for PMS Evolutionary Models

The directly measured dynamical mass for S1A ($\nu=5$9) lies $M_\odot$025–50% below the mass inferred from contemporary PMS evolutionary models (PISA, YALE-Potsdam, Palla-Stahler, Y$M_\odot$1, PARSEC) at its observed HRD location and luminosity. Neither non-rotating nor rotating isochrone grids reconcile the observed discrepancy. No standard models incorporating magnetic fields or modest rotation resolve this offset. This result suggests a systematic overestimate of mass (or luminosity) by intermediate-mass PMS tracks for stars at ages $M_\odot$21 Myr. In contrast, masses for EC 95A and EC 95B ($M_\odot$32 $M_\odot$4) are in agreement with theoretical models, providing critical validation in the low- to intermediate-mass range.

For EC 95, the architecture—with similar inclination angles for inner and outer orbits but high eccentricities—suggests disk fragmentation as the most likely formation channel, though near-periastron passages of EC 95C may affect long-term dynamical stability.

6. Data Analysis, Validation, and Systematics

All positional and flux measurements for DYNAMO-VLBA are subject to rigorous internal cross-checking and error estimation. Calibration errors, astrometric error floors (30–100 $M_\odot$5as), and inter-epoch consistency are formally accounted for. For orbital fitting, both least squares (MPFIT) and MCMC-based Orbitize! (when appropriate) are used to sample the posterior space, avoiding local minima.

Revisions to previous mass estimates (e.g., S1A) result from correction of astrometric outliers, additional high-S/N detections (S1B), and uniform self-calibration across epochs. Uncertainties are propagated via analytic derivatives or posterior sampling.

7. Future Directions and Planned Expansions

The DYNAMO-VLBA project will extend its sample to $M_\odot$620 additional radio-loud binaries and higher-order multiples in other star-forming regions (e.g., Taurus, Orion, Perseus). Multi-frequency VLBA campaigns are planned to probe correlations of radio and X-ray emission, and to infer the size and morphology of magnetospheres. Continued temporal baselines will enable complete mapping of outer orbits in hierarchical systems, essential for understanding formation and dynamical evolution. Synergistic observations with JWST and ALMA are anticipated to combine astrometric mass constraints with resolved spectroscopy and disk properties, ultimately refining physical models of early stellar evolution and star formation.

DYNAMO-VLBA thus establishes a robust empirical framework for the calibration and potential revision of PMS evolutionary models in both low- and intermediate-mass regimes, and for the study of the dynamical architectures of young multiple stellar systems.