Physics, Earth & Space Icon Physics, Earth & Space

A Crisis in Cosmology?


Cosmology joined the “precision science” club in the 1990s. In that decade the COBE probe measured the sought-after ripples in the cosmic microwave background radiation (CMBR) for the first time. Also, astronomers used Hubble Space Telescope observations of cepheid variable stars in nearby galaxies to calibrate the Type Ia supernova standard candle. Then, with this new tool astronomers measured the current value of the Hubble constant, H0 (pronounced H-not), much more precisely and to greater distances than previously possible. H0 is a measure of the present expansion rate of the universe; “local distance ladder” measurements require distances (thus the standard candles) and spectroscopic Doppler shifts of the host galaxies. For local galaxies, it is the slope of Doppler velocity plotted against distance. It is often reported in units of km/s/Mpc (which stands for kilometers per second per megaparsec). 

At the same time, astronomers discovered that the new data required an additional term in Einstein’s general relativity field equations, a cosmological constant (parameterized as lambda). It is responsible for acceleration of the receding galaxies. The mysterious energy responsible for the cosmological constant is called dark energy. Its discoverers were awarded the Nobel Prize for physics in 2011.

As the quality and quantity of observations have improved, cosmologists have continued to rule out competitors to the so-called standard model of cosmology, lambda-CDM (lambda-cold dark matter). It is the theoretical basis for the Big Bang theory, the idea that the universe began in a hot compact state. For a brief summary of the history of the Big Bang theory see here and here. The lambda-CDM model rests on a multitude of diverse and precise observations. At this time, it has no serious competitors.

A Crisis Emerges

Following COBE two additional probes have measured the CMBR with ever greater precision, WMAP and Planck. Every time cosmologists improve their measurements of the CMBR, they constrain the properties of the universe better. In particular, H0 can be constrained from the CMBR in combination with observations relating to the present (energy equivalent) matter density fraction, Ωm. One such observation is the primordial D (deuterium) abundance, which when combined with a Big Bang nucleosynthesis model, gives Ωm. Another approach is to measure how galaxies are distributed on large scales (baryon acoustic oscillations, or BAO) in combination with the CMBR. These estimates of H0 depend how well the models account for conditions in the early universe. Other methods used to determine H0 are described here and here. These estimates from highly diverse methods average near 70 km/s/Mpc (more on this below).

Ten years ago things seemed to be going well. H0 estimates appeared to be converging. However, a simple average hides important details. Each H0 estimate has an associated uncertainty (often in two parts, systematic and random). As the uncertainties in H0 became smaller, values determined from different methods failed to converge to a single consistent value within their respective uncertainties. The first inkling of a problem occurred in 2013 when the initial Planck value for H0 was published, 67.3 ± 1.2 km/s/Mpc. In 2016 a redetermination of the “local” value of H0 based on recalibrations of the cepheid Leavitt Law and the Type I a supernovae resulted in a value of 73.2 ± 1.7 km/s/Mpc. The difference was greater than 3 sigma (standard deviations); there is less than 1 percent probability that these estimates really are the same although astronomers measured them to be different. 

A Growing “Tension”

Cosmologists officially declared the growing Hubble constant “tension” to be a “crisis in cosmology” at the April 2018 meeting of the American Physical Society in Columbus, Ohio. Since then, the situation has only become worse. This year, the difference in H0 determinations has surpassed 5 sigma (see here and here).

There are two main ways to resolve the H0 crisis. First, look for unaccounted for systematic errors in the local approach. There are multiple rungs in the local distance ladder, and a small error in any one of them can affect the final answer. At this point, this seems unlikely, given the close agreement of the multiple diverse local methods employed. Second, adjust the standard lambda-CDM model or even introduce new physics that is important at early times. The most reasonable tweak would be to increase the number of neutrino species beyond three, but all the experimental evidence points to only three species. Many are rooting for new physics; it’s exciting. The last time there was a problem with H0, cosmologists discovered dark energy.

Not surprisingly, observational and theoretical cosmologists have been busy in recent months proposing solutions to the H0 crisis. Hardly a week passes that a new paper is not published on the astro-ph preprint server about it; sometimes two or three papers are posted on the same day. 

A Solution on the Horizon?

Before you throw out the lambda-CDM model altogether, consider that the mean age of the oldest stars (in globular clusters) is 13.2 ± 0.4 billion years. Within the context of the lambda-CDM model, this translates to a value for H0 of 71.0 ± 2.8 km/s/Mpc. This just happens to fall almost in the middle of the early and late H0 measurements. It tells us that the value of H0 astronomers are measuring are in the right ballpark for the lambda-CDM model; the eventual corrections will be in the neighborhood of 2 or 3 km/s/Mpc.

Some progress is being made at erasing the gap between the early- and late-universe H0 estimates. I’ll just mention a couple of papers to give a flavor of the research being done. One study seeks to recalibrate the cepheid Leavitt law using parallaxes from the second data release of the Gaia astrometric space mission. They calculated a local value for H0 of 69 ± 2 km/s/Mpc, which is consistent with the early universe value. We should know within the next couple of years whether this holds up with the next Gaia data release.

Another study argues that galaxies are distributed more inhomogeneously “locally” (out to about 200 Mpc) than assumed. In particular, we might be within a “Local Hole,” an underdensity in the galaxy distribution. This, possibly combined with a “nearby” supercluster of galaxies and galaxy clusters (the Great Attractor), is causing peculiar motions directed away from us. Ignoring this motion leads to a local value of H0 that is too large by 2 to 3 percent, about half of what is needed to close the gap.

It’s too bad biologists are not as open about crises in their theories. It’s the sign of a healthy science that problems are openly acknowledged and discussed by its practitioners.

Photo credit: The heart of the Milky Way, captured by Spitzer Telescope, via NASA, JPL-Caltech, Susan Stolovy (SSC/Caltech), et al.