An observed number of gravitational lenses exceeded the simulation forecast by an order of magnitude
Physicists estimated the number of small-scale gravitational lensing events in galactic clusters based on observations and then compared it with the result of computer simulations. It turned out that the simulation predicts an order of magnitude lower number of such objects — the result may call into question the generally accepted ideas about the properties of dark matter, or speak about missing systematic errors in simulations. The article is published in the journal Science.
According to General relativity — the most successful theory of gravity (in terms of describing experimental data) to date — the presence of a massive object distorts space-time in its vicinity. Due to this, rather heavy bodies (in particular, galaxies and their clusters) that are in the path of propagation of electromagnetic radiation are able to noticeably change its direction, bending the rays around themselves. This effect is commonly called strong gravitational lensing, and the bodies themselves are called gravitational lenses. An observer for whom the light source is blocked by a gravitational lens will generally observe its multiplied curved images — their shape, number, and position will be determined not only by the position of the background source relative to the lens and observer but also by the mass distribution in the lens itself. Due to this, gravitational lensing events help scientists in detecting and investigating massive structures — including formations of dark matter inaccessible to direct observation in electromagnetic radiation. Physicists from
Denmark, Italy, the Netherlands and the United States under the leadership of Massimo Benedetti from the National Institute of astrophysics of Italy has examined one of the possible types of strong gravitational lensing — the cases when as the background source and lens are distant galaxies, and galaxy-lens is part of a cluster of galaxies — i.e. it is a small lens (divided parts of the image in the order of arcseconds) on the background of a much larger (divided in tens of angular seconds).
Scientists turned to observations of 11 galactic clusters in the redshift range of 0.234–0.587, images of which were obtained by the Hubble space telescope in 2010-2016 during the programs CLASH (Cluster Lensing and Supernova Survey with Hubble) and HFF (Hubble Frontier Fields). In addition, the researchers used spectroscopy data from the ground-based VLT (Very Large Telescope) complex this allowed us to establish the mass distribution inside the cluster in more detail (based on the nature of stellar movement).
Based on this data set, physicists calculated how much of the total angular area of small-scale gravitational lenses is equal to the total angular area of a characteristic region within which background sources are subjected to strong lensing. Using this number, the authors estimated the probability of detecting a small-scale lensing event at a given redshift and, as a result, calculated the estimated number of visible events in one cluster, assuming the known number of lensed background galaxies (based on Hubble Ultra-Deep Field data - a set of observations of distant galaxies by Hubble).
The researchers calculated the same probability of detecting small-scale lensing for purely theoretical reasons. The scientists modelled galactic clusters of the same mass and size as those in the observational sample using hydrodynamic simulations based on the current standard cosmological model, taking into account gas cooling, star formation, and energy input from supernovae and accreting supermassive black holes.
As a result, scientists have found that the theoretical predictions of the probability of small-scale gravitational lensing deviate from the experiment by about an order of magnitude — in fact, we see many more events than the simulations predict.
This discrepancy could not be compensated for even by excluding from simulations the feedback from supermassive black holes, which suppresses star formation and, by making the internal mass distribution smoother, weakens gravitational lenses. The experimental calculations turned out to be self-consistent — for example, for the MACSJ1206 cluster, the expected number of events was about three, which coincided with the observed number.
The problem, therefore, relates specifically to theoretical calculations. The authors believe that the reason for the discrepancy with the observations could be both methodological errors in the numerical modelling procedure itself, and incorrect initial assumptions of the cosmological model concerning the properties of dark matter and its interaction with ordinary matter.
Gravitational lenses make it possible to study not only the structure of themselves but also the background objects. In recent years, we have described how one such lens allowed us to see four moments of the life of a single supernova at once, and another lens helped x-ray observations of the galaxy from the young Universe.
Photo: Massimo Meneghetti et al. / Science, 2020