A New Clock for Globular Clusters: Brown Dwarfs That Never Fuse Hydrogen

In the crowded heart of an ancient globular cluster, most stars shine steadily for billions of years. Brown dwarfs do not. They are substellar objects, below the hydrogen-burning limit of about 0.07 to 0.09 solar masses, and they spend their lives cooling. That simple fact may offer a new way to date some of the Milky Way's oldest objects.

A clock below the hydrogen-burning limit

Globular clusters are massive, compact swarms of stars. Some 150 are known in the Milky Way. Many are older than 12 billion years, and their ages help reconstruct how the Galaxy assembled from smaller pieces. But measuring those ages precisely has been stubborn.

The standard method looks at the main sequence turnoff, the point in a cluster's color-magnitude diagram where stars stop fusing hydrogen in their cores and begin evolving toward becoming red giants. The position of that turnoff depends on age, so it can act as a clock. The trouble is that it also depends on distance, interstellar dust, chemical composition, and how models treat convection, the churning motion that moves heat inside stars. Different analyses produce precise but discordant ages, hinting at systematic errors larger than 1 billion years.

Alternative clocks exist. White dwarf cooling sequences and detached eclipsing binary stars have yielded ages for a handful of clusters, but each method has its own limitations and distinct systematics. The question is whether astronomers can add another clock whose errors are largely independent.

Roman Gerasimov of the University of Notre Dame proposes that brown dwarfs can serve as that clock. The paper, accepted for publication in Astrophysics and Space Science and posted as a preprint, describes a method for deriving globular cluster ages from JWST observations of cluster members near and below the hydrogen-burning limit.

The logic is tied to a feature called the stellar/substellar gap. In a young cluster, brown dwarfs and the lowest-mass stars cool and fade in similar ways. After roughly 100 million years, however, the low-mass stars settle onto the main sequence and stabilize, while brown dwarfs keep cooling indefinitely. A gap opens in the luminosity function, the number of objects at each brightness. The size and depth of that gap depend on the cluster's age.

Seeing this gap is hard. Brown dwarfs are faint and radiate most strongly in infrared light. Before JWST, globular cluster brown dwarfs were effectively invisible. So far, only one globular cluster, NGC 6397, has an age estimate from its brown dwarf cooling sequence: 13.4 plus or minus 3.3 billion years, based on just three brown dwarf members.

Gerasimov's study is not a new observation of a cluster. It is a forecast of what JWST could do. He simulated deep observations of the nearby, massive cluster 47 Tucanae using NIRCam, JWST's near-infrared camera, in two ultra-wide filters, F150W2 and F322W2. The simulations included realistic measurement errors and completeness, the loss of faint objects near the detection limit.

The method is histogram-free and likelihood-based. Instead of sorting stars into brightness bins, which can smear errors and depend on bin choices, it compares the measured magnitudes directly with model luminosity functions. Proper motion measurements from two epochs of imaging, taken two or five years apart, help separate true cluster members from foreground stars and background galaxies.

Simulating 47 Tuc and NGC 6397

For a single JWST field similar to earlier observations of 47 Tuc, the expected formal age errors range from about 0.3 to 1 billion years, depending on exposure time and the time between observations. Formal errors here mean the random uncertainties from the simulated measurements and analysis, not the full systematic error budget. Two 10-hour exposures separated by five years gave the smallest single-field error, about 0.3 billion years. Two 1-hour exposures five years apart gave about 0.5 billion years and appeared to be the most efficient use of telescope time.

With three or four non-overlapping fields, formal errors as low as about 0.2 billion years are within reach for nearby clusters, according to the simulations. That is comparable to the smallest formal errors quoted for main-sequence turnoff ages, though the paper emphasizes that it is not the whole story.

The study also provides a lookup table for 30 nearby globular clusters. It estimates how many age-sensitive objects, mostly brown dwarfs, a given observing strategy might reveal and what formal age error that sample would imply. For NGC 6397, for example, a field comparable to one previous JWST observation might yield about 56 such objects and a formal error near 0.95 billion years. Reaching about 0.3 billion years would require observing roughly seven similar fields.

Systematics that dominate the error budget

As with other globular cluster dating methods, real-world accuracy is limited by systematic effects. The new study evaluates several.

Globular clusters are not chemically uniform. They host multiple populations with varying abundances of light elements, especially oxygen. Those variations change the infrared colors and the inferred luminosity function of cool brown dwarfs. Multiple populations alone can inflate the formal error by about a factor of two. If the chemical spread changes with stellar mass, as recent JWST observations of 47 Tuc may suggest, simulations produce an additional bias toward older ages of about 0.45 billion years and can inflate the spread by about a factor of three.

Uncertainties in brown dwarf cooling models matter too. A +0.2 dex error in the assumed metallicity, the abundance of elements heavier than helium, could inflate the formal error by up to a factor of five and shift the age older by about 0.4 billion years. Unresolved binary systems are less important for clusters with binary fractions below 10 percent, but clusters with fractions above 20 percent could suffer age errors exceeding 0.2 billion years, reaching as high as 1.2 billion years in a worst-case simulation. Errors in estimating completeness, the fraction of faint cluster members detected, also add scatter.

Adding these effects, a nominal formal error of 0.3 billion years could translate into an overall uncertainty of roughly 1.5 to 2 billion years. That is comparable to the systematic error budgets of the main sequence turnoff method. The advantage is that the two methods' errors should be largely independent, making brown dwarf dating a useful consistency check rather than a replacement.

The oldest globular clusters preserve a record of the Milky Way's assembly, including clusters formed in place and clusters accreted from smaller galaxies. Their ages help astronomers interpret that record, but only if the clocks agree. The brown dwarf method is not yet ready to set the standard. It needs larger samples, more reliable brown dwarf cooling models, and careful treatment of multiple populations and binaries. The 5-year baselines considered here are already comparable to the telescope's expected remaining lifetime as quoted in the paper, and the faintest members require long exposures.

Still, the idea turns a limitation into an opportunity. The same faintness that once hid brown dwarfs makes them sensitive age markers. If future surveys can identify enough of these cooling embers, the Milky Way's oldest star clusters may get a new, independent timeline, written in objects that never quite became stars.