76. The James Webb Space Telescope, and the early Universe

(by Moritz Haslbauer & Pavel Kroupa, Saturday 29th October 2022)

Over more than three decades the Hubble Space Telescope (HST) has improved our understanding of the Universe through varied observations of objects in our Solar system and in the very distant Universe. This year, Hubble’s successor, the James Webb Space Telescope (JWST) named after the second  Administrator of NASA, James E. Webb, has started its observations. Beside beautiful images of, for example the Carina Nebular, Stephan’s Quintet, and the Southern Ring Nebula, the first observations by JWST delivered scientific results which puzzled astronomers and astrophysicists: Galaxies in the very early Universe are much more massive than predicted by the standard model of cosmology (ΛCDM). But has the JWST indeed already falsified dark-matter-driven galaxy formation?

Let us start from the beginning: After more than twenty years of development, the JWST was finally launched on 25 December 2021 and sent to the Sun–Earth L₂ Lagrange point, which is a special place at which the gravitational force of the Sun and Earth and the centrifugal forces balance. Thus, JWST has a relatively stable orbit and requires only a few orbit corrections once for a time. The JWST operates in the near-infrared regime, which is mostly absorbed by the Earth’s atmosphere and thus not accessible with earthbound telescopes. An illustration of the JWST with its 6.5 meter mirror is shown in Figure 1.

Figure 1: The James Webb Space Telescope has a primary mirror with a diameter of 6.5 meter and is located at the Earth-Sun L₂ Lagrange point. The silver layers are sunshields which protect the telescope from radiation emitted by the Sun. Credits: NASA

After a phase of testing, which included e.g. the alignment of the mirrors, the first image of the JWST was presented to the world wide public on 11 July 2022 during a press conference. The image is called “Webb’s First Deep Field” (Figure 2) and shows the deepest and sharpest view of the Universe revealing galaxies just a few 100 Myr after the Big Bang. The image also illustrates the capability of the JWST: While multiple weeks were required to take the “Hubble Deep Field” by the HST, the integration time of “Webb’s First Deep Field” took only 12.5 hours!

Figure 2: The image shows the “Webb’s First Deep Field” and is the first image taken by the JWST. Galaxies in front are part of the galaxy cluster SMACS J0723.3–7327. The high mass of this galaxy cluster deflects the light of background galaxies (reddish objects) making them to appear disturbed (gravitational lensing effect). Credits: NASA, ESA, CSA, STScI

Thus, the JWST gives new scientific insights to the evolutionary stage of galaxies in the very early Universe. Research teams (e.g. Adams et al. 2022, Atek et al. 2022, Labbe et al. 2022, Naidu et al. 2022a, Naidu et al. 2022b) analysed the first data acquired by JWST and found several massive objects with stellar masses of M_* > 10⁹ M_☉ just 200-500 Myr (corresponding to redshifts of z ≈ 10-17) after the Big Bang; and some of them have stellar masses even higher than our own Milky Way galaxy! This is striking because the Milky Way has build up a stellar mass content of M_* ≈ 5⋅10¹⁰ M_☉ over 12 Gyr. Up to now, the JWST only measured the redshifts based on photometric data, but upcoming measurements of the spectra are required to obtain the real ‘spectroscopic’ redshifts. Therefore, we refer to these objects as ‘galaxy candidates’ in the present blog post.

Figure 3: The diagram shows the most massive galaxies in terms of the stellar mass at different redshifts (coloured points, as the Universe ages, the redshift z decreases to the left) formed in the ΛCDM simulations. The observed galaxy candidates depicted as grey and black errorbars are more than 10 times more massive than the most massive galaxies formed in the ΛCDM simulations (coloured points). GN-z11 is a spectroscopically confirmed galaxy detected with the Hubble Space Telescope (Oesch et al. 2016). Naidu et al. 2022a updated the physical properties of the galaxy candidates GL-z11 (now labeled as GL-z10 in Naidu et al. 2022a) and GL-z13 (now GL-z12) in the accepted version of their paper. Their updated redshifts and stellar masses are visualized as grey errorbars in the figure. Unfortunately, they authors updated these values after our paper got accepted such that we were not able to include them in the analysis of our paper. However, the updated values do not change the conclusion of our paper in any way as can be seen by the grey error bars (GL-z10 and GL-z12). Credits: Modified version of Fig. 2 of Haslbauer et al. 2022b

In our recent publication titled “Has the JWST already falsified dark-matter-driven galaxy formation?“, we investigated if the observed galaxy candidates by the JWST are consistent with the standard model of cosmology assuming that these galaxy candidates are indeed located at such high redshifts. Therefore, we used the state-of-the-art IllustrisTNG and EAGLE simulations and looked for the most-massive galaxies formed at different redshifts. Galaxies formed in the standard model of cosmology are by more than one order of magnitude less massive than the observed galaxy candidates as shown in Figure 3. Similar results have been also obtained e.g. by Boylan-Kolchnin et al. 2022 or Lovell et al. 2022. Essentially, the results show that the dark-matter-based cosmological models need much more time for galaxies to grow through mergers and accretion than in the observed real Universe. The observational data would thus rule-out these dark-matter based models. But is there perhaps a possible way out of this conclusion?

Possible reasons for the discrepancy:

Several reasons for the discrepancy between the observed and simulated stellar mass buildup of galaxies have been proposed. First of all, it could be that there is an error in the calibration of the JWST and that the measured stellar masses and redshifts of these objects are wrong. Upcoming followup observations are now necessary to address this possibility. Secondly, it has been argued that these galaxy candidates are surrounded by dust which make them to appear redder which increases therewith the photometric redshift. A highly interesting object is the galaxy candidate CEERS-1749 identified by Naidu et al. 2022b, which could be either an extremely high-redshift galaxy located at redshift z = 17 (so just 230 Myr after the Big Bang!) or a very dusty galaxy at redshift z = 5 when the Universe was about 1.2 Gyr old. The authors refer to that object as Schrödinger’s galaxy candidate as an (humorous) allusion to Schrödinger’s cat. But, if the galaxies feign a high (photometric) redshift because they are in truth highly obscured by dust, then a new problem arises, namely, what made the dust? The galaxies are, even at z=5, too young to have been able to produce enough dust to enshroud them, as not enough stars have yet died. In this context, the results by Rodighiero et al. 2022 are of interest, because they found galaxies with M_* = 10⁹ – 10¹¹ M_☉ in the redshift range of z = 8-13 having a high dust content questioning therewith the ΛCDM model.

There is also another possibility which we qualitatively discussed in our paper. Telescopes cannot directly measure the stellar mass content of a galaxy but only measure the emitted light of the stellar population. The measured luminosity therefore needs to be converted to a stellar mass, which requires knowledge of the stellar mass-to-light ratio of a galaxy. This conversion factor depends on the mass distribution of stars and is called the initial mass function (IMF). In the analysis the researchers assumed an invariant canonical IMF, meaning that the mass distribution of stars is always the same independent of the global properties of the galaxy. However, recent observations have shown that the IMF changes depending on the physical properties of the star-forming region. For example, galaxies with a low metallicity and high star formation rate are described by a so-called ‘top-heavy galaxy-wide IMF’ because they host more massive stars than predicted by an invariant canonical IMF.

But beware: one cannot just arbitrarily invent some galaxy-wide IMF form to solve a problem. The galaxy-wide IMF must, in reality, be calculated carefully from the way stars from throughout a galaxy. This is done in the “Integrated Galaxy Initial Mass Function” (IGIMF) theory (as outlined by Kroupa et al. 2013, Jerabkova et al. 2018, Yan et al. 2021, Wirth et al. 2022) by summing up all star-forming regions in a galaxy that are contributing stars at a given time. Thus one adds up all the metallicity and molecular-cloud density dependent stellar IMFs in all star-forming events (embedded star clusters) to obtain the galaxy-wide IMF of all stars just forming in the galaxy.

Adopting this variation of the galaxy-wide IMF has interesting implications: Massive stars produce much more UV radiation than low massive stars. If we now measure the UV luminosity of a galaxy and wrongly assume that the IMF is invariant, we would overestimate the stellar mass content of the galaxy. Figure 4 shows that the stellar masses of the galaxy candidates would indeed be smaller by a factor of 5-10 times than for an invariant canonical galaxy-wide IMF. A more sophisticated analysis is still required in order to test if a varying IMF can resolve the tension because the stellar masses that form are affected by the intense feedback from the forming massive stars, while the existing simulations of galaxy formation assume the galaxy-wide IMF to not change. In reality, there is thus a much larger degree of self-regulation of the growth of galaxies: due to the initially small metallicity the early IMF would be top-heavy. The large luminosity of the stars would heat the gas and reduce the star formation rate, such that the real galaxies would end up being less massive.

Figure 4: The solid coloured data points show the stellar masses of the galaxy candidates assuming a non-constant galaxy-wide IMF calculated self-consistently from the assumed metallicity and observed luminosity. The calculation uses a stellar IMF which depends on the metallicity and gas density of the star-forming molecular clouds, and assumes a constant galaxy-wide SFR over cosmic time. The star-like error bars show the measured stellar masses of the galaxies assuming an invariant canonical galaxy-wide IMF. The coloured squares show the most massive galaxies found in the simulations of a ΛCDM universe and assuming an invariant canonical galaxy-wide IMF. Credits: Haslbauer et al. 2022b

Maybe indications of a different cosmology?

Finally, the existence of massive galaxies just 200-500 Myr after the Big Bang implies that structure formation is much more enhanced at high z, or that the Universe is much older than predicted by the standard model of cosmology. An enhanced growth of structure formation would be in line with the existence of the massive El Gordo Galaxy Cluster or with the huge KBC void observed in the “local” (400 Mpc radius) Universe – both individually falsify ΛCDM with more than 5 sigma (Asencio et al. 2021, Haslbauer et al. 2020). As discussed in our previous blog posts (e.g. 52 or 54), an enhanced growth of structure is expected in Milgromian Dynamics (MOND) because of a stronger self-gravity. But simulations of such MOND-based cosmological models need to be performed to find out if the observed massive galaxies in the very early Universe can arise in such models. Hydrodynamical cosmological simulations conducted in the framework of MOND in order to address these questions are currently under construction in the Bonn-Prague SPODYR research group. Naturally, this undertaking is difficult, because the community generally does not allow one to get research money for this enterprise. But a thick skin, perseverance, a capacity to be sufficiently confrontational (apparently needed unfortunately), are slowly but steadily leading to advances. So stay tuned!

Final remark:

But let us return to our question raised at the beginning: Has JWST already falsified dark-matter-driven galaxy formation? At the moment, this question cannot be conclusively answered. That is, we cannot now conclude that the JWST observations of the early Universe have falsified the standard, Newton/Einstein-based models with more than 5sigma confidence. Measurements of the spectroscopic redshifts are required and it must be ensured that the JWST is correctly calibrated. But it is true that the observational data, wherever one looks (as documented in The Dark Matter Crisis), appear to always show in the same direction, namely, hat these dark-matter based models of structure formation and of galaxies do not work.

Addendum (added on 30th of October as a result of communications with Andrea Ferrara):

Andrea Ferrara points out (private communication) that if the observed galaxies by Naidu et al. (2022a) are in actuality dust free, then their blue spectra would masquerade as a more heavy population, as argued by Ferrara et al. (2022). Our modelling already assumed dust-free populations, so this is OK. Our models are very blue because they are dust-free and are based on a top-heavy galaxy-wide IMF. Dust free populations in the high-z Universe may be realistic due to radiatively-driven outflows (Ziparo et al. 2022). UV spectra of the high-redshift galaxies that have extremely blue UV spectral slopes (beta = -2.6, Ziparo et al. 2022) can be a natural outcome of the top-heavy IMF, especially if the light emission is dominated by the central hypermassive star-burst cluster (fig.8 in Jerabkova et al. 2017). Such “Jerabkova objects” evolve into SMBHs within 100-300 Myr as their hosting galaxy keeps forming on the down-sizing time-scale (Kroupa et al. 2020; Eappen et al. 2022). 

---

In The Dark Matter Crisis by Moritz Haslbauer, Marcel Pawlowski and Pavel Kroupa. A listing of contents of all contributions is available here.