On the Reported Death of the MACHO Era

TL;DR

This work tests MACHO dark matter scenarios by obtaining radial velocities for four wide halo binary candidates from the CG04 sample. It confirms three of the four widest binaries are real, including the widest, while identifying a spurious interloper among them. The authors show that the orbits of these binaries imply time-averaged dark matter densities along their paths are significantly lower than the local density used in earlier analyses, weakening the resulting MACHO constraints. They emphasize that the current wide-binary constraints are highly sample-dependent and advocate for a much larger, cleaner dataset (e.g., Gaia) to robustly probe halo dark matter structure.

Abstract

We present radial velocity measurements of four wide halo binary candidates from the sample in Chaname & Gould (2004; CG04) which, to date, is the only sample containing a large number of such candidates. The four candidates that we have observed have projected separations >0.1 pc, and include the two widest binaries from the sample, with separations of 0.45 and 1.1 pc. We confirm that three of the four CG04 candidates are genuine, including the one with the largest separation. The fourth candidate, however, is spurious at the 5-sigma level. In the light of these measurements we re-examine the implications for MACHO models of the Galactic halo. Our analysis casts doubt on what MACHO constraints can be drawn from the existing sample of wide halo binaries.

Figures (4)

• Figure 1: Spectra for the four candidate binary systems we have observed. The top three panels show data from our WHT observations, while the bottom panel shows the data for the pair NLTT 16394/16407 taken with the Magellan telescope. In each panel, the spectra are centred on the Ca II triplet region. The spectra for each member of a candidate binary are plotted together. The second component is shifted slightly upwards for clarity.
• Figure 2: A comparison of the predicted observable angular separation function for a number of models. See text for details. The observed distribution, from the CG04 homogeneous sample less the spurious pair, is also shown with associated Poisson errors.
• Figure 3: Confidence regions for the MACHO mass $M_{\rm p}$ and MACHO halo fraction. For comparison to past work we show the $2\sigma$ joint confidence levels as defined in Yoo04 and using the standard definition, respectively, when the whole CG04 homogeneous sample is included. We also show the updated $2\sigma$ confidence levels omitting the spurious candidate binary. The omission of this object eases the constraints on MACHOs; the window increases to $\approx 30-500$ M$_{\odot}$. In addition, the effect on the constraints of omitting the widest binary in CG04 is shown at the 90% confidence level: the constraints at the $2\sigma$ level vanish. The regions of parameter space shaded in grey are ruled out at the $2\sigma$ level by binaries and microlensing data -- an upper limit on the MACHO mass and halo fraction from disk kinematics is also shown. We stress that the constraints from the binaries are based on the assumption that the time-averaged dark matter density experienced by each binary is the local halo density at the position of the Sun -- the actual Galactic orbits of the confirmed wide binaries suggest much lower time-averaged dark matter densities. See text for a detailed discussion.
• Figure 4: Orbits over 10 Gyrs for the 3 wide binaries that we confirmed and wide binary NLTT 39456/39457. The Milky Way Mass model 1 of DehnenMW is assumed and for clarity we have flipped the sign of R for NLTT 15501/15509.