My research

My research interests mainly revolve around galaxy formation and evolution in the early Universe: how the first stars and galaxies were formed, how they subsequently started forming elements heavier than hydrogen and helium, and how they brought about the Reionization of the Universe, a major phase transition that turned the neutral Universe back to a (mostly) ionised state.

In my current research, I am involved in multiple different projects with various collaborators at the Kavli Institute for Cosmology. I mainly work on observational data of high-redshift star-forming galaxies from several major observatories.

Together with Dr Renske Smit and Prof Roberto Maiolino, and others, I am involved with the James Webb Space Telescope (JWST) Guaranteed Time Observations (GTO) collaboration of its NIRSpec instrument. In particular, I am an associate member of JADES (the JWST Advanced Deep Extragalactic Survey), an ambitious joint JWST NIRCam and NIRSpec GTO effort that is conducting the largest extragalactic survey currently planned for JWST, having been allocated ~800 hours of observing time.

We also work with Atacama Large Millimeter/submillimeter Array (ALMA), the Hubble Space Telescope (HST) and the Very Large Telescope (VLT; in particular the MUSE and X-shooter instruments). By combining observations from these facilities and comparing them to theoretical models, we are exploring the physical conditions of star formation in the very early Universe, and its implications for the subsequent evolution of galaxies up to the present day.

With Dr Ewald Puchwein, Dr Girish Kulkarni, Dr Renske Smit, Prof Martin Haehnelt, and Lewis Weinberger, I have further been involved in analysing the Sherwood simulation suite, a set of large-scale cosmological hydrodynamical simulations aiming to accurately model the intergalactic medium. Specifically, we are working on making theoretical predictions of the brightness of this vast reservoir of diffuse gas in the principle emission line of hydrogen, Lyman-alpha.

Below follow few short excerpts summarising work I am involved in.

NIRCam images indicating four galaxies (left) with spectra taken by NIRSpec (right). Credit: ESA.

In December 2022, JADES announced the discovery (with NIRCam) and subsequent spectroscopic confirmation (with NIRSpec) of four of the most distant galaxies, including the two most distant galaxies to date. These results were highlighted on the ESA website of JWST.

New measurements with JWST have discovered and unambiguously confirmed the most distant galaxies known to date, representing a major breakthrough in understanding our cosmic origins and promising a bright future for further discoveries with JWST.

Previously, the first imaging surveys taken by JWST had revealed candidate galaxies seen when the Universe was less than 2% of its current age. Identifying and characterising these distant galaxies is essential to our understanding of galaxy formation and, more broadly, of the Universe as a whole. Preliminary analyses have indicated the abundance of these candidates may violate prevailing cosmological models of galaxy formation. Without spectroscopic confirmation of their distances, however, these findings remain speculative.

New results from the JADES collaboration have now confirmed the nature of four of the most distant galaxies, including the two most distant to date. The NIRSpec instrument was able to provide spectroscopic confirmation of four galaxies initially identified in one of the deepest NIRCam imaging survey so far (see figure on the left). The analysis is presented in two companion papers, Curtis-Lake, E., et al., 2022, arXiv:2212.04568 and Robertson, B., et al., 2022, arXiv:2212.04480.

The following is a short description of our findings in a joint study of the [O III] 88 μm and [C II] 158 μm emission lines with ALMA, conducted in collaboration with Dr Renske Smit, Prof Roberto Maiolino, and others.

The title of this work, Dual constraints with ALMA, points to its two-faced nature, dealing with two types of spectral features: line and dust-continuum emission. It’s also a tale of two observatories, ALMA and HST, looking back about 13 billion years in time.

We used the powerful ALMA observatory to observe the [O III] 88 μm emission line in five bright 𝑧 ∼ 7 Lyman-break galaxies. These systems were first spectroscopically confirmed by ALMA through the [C II] 158 μm line, setting them apart from recent [O III] detections where Lyman-α was used. With the Hubble Space Telescope, we obtained new images of the rest-frame ultraviolet, revealing their young, star-forming regions in Hubble’s trademark ultra-high resolution. The [O III] 88 μm emission generally traces these regions well, but not always perfectly. We find [C II] to be more extended (see top figure on the right).

A non-detection of [N II] 205 μm in one source shows (via the [C II]/[N II] ratio) that for typical physical conditions, the [C II] emission can likely be traced to a neutral medium, as the gas in photodissociation or X-ray dominated regions (PDRs/XDRs).

The [O III]/[C II] ratio – seemingly elevated in high-redshift galaxies – has been hotly debated in recent years. We find a tentative anti-correlation between this ratio and the degree of dust obscuration, both on global and local scales, as well as in nearby galaxies. However, when we start to look at the [O III] luminosity in detail, Cloudy models struggle to reproduce its strength, unless we assume an α/Fe enhancement and a near-solar nebular oxygen abundance.

There’s more surprises when we look at the dust continuum. Constraints at 90 μm and 160 μm allow us to study (currently poorly understood) dust properties such as the temperature. To do so, we present a newly developed far-infrared spectral energy distribution fitting routine, MERCURIUS. Humbly named after the messenger of the gods, it tells us about the presence of cold dust in one system (bottom figure on the right), implying a large dust mass that would require a high yield of metals and efficient subsequent conversion into dust grains.

This study was published as Witstok, J., et al., 2022, MNRAS, 515, 1751.

Detections of the [C II] 158 and [O III] 88 μm lines. Top row: contour images of the [C II] and [O III] lines overlaid on a background of HST rest-frame UV images. Bottom row: S/N-weighted line spectra extracted from a 2σ region in a naturally weighted image integrated over the FWHM. Coloured channels, which were used to produce a subsequent moment-zero map for measuring the total line flux, indicate the line detection. The significance of each line detection is shown in the legend. Figure 2 from Witstok et al. (2022).
Dust peak temperature, Tpeak, as a function of cosmic time. Several peak temperatures derived from measurements presented in this work under the self-consistent general opacity model are drawn: for UVISTA-Z-019, we show its median and 68 per cent confidence range (largely invariant under the chosen opacity model), while for COS-3018555981 an upper limit (95 per cent confidence) is shown in addition. Three sources are highlighted with labels, having either a notably low (J1217−0208), high (MACS0416-Y1), or accurately constrained dust temperature (A1689-zD1). The grey dashed line shows the CMB temperature over cosmic time. The light-red area above indicates the region where CMB heating has a significant (greater than 1 per cent) effect compared to z = 0. Figure 4 from Witstok et al. (2022).
Top left: false-colour image of the foreground cluster and the two arcs (A1–A3) of the background galaxy RCS0224z5 observed with X-shooter and SINFONI. Other sections show the detected emission lines. Figure 1 from Witstok et al. (2021b).
Measurements of the metallicity offset from the Fundamental Metallicity Relation (FMR), which predicts the amount of heavy elements in galaxies that have reached an equilibrium state. Our results show a tentative departure from the predicted value for RCS0224z5, as well as for several other high-redshift galaxies from the literature, potentially indicating a recent inflow of pristine (metal-poor) gas. Figure 5 from Witstok et al. (2021b).

This is a short article describing our work on X-shooter and SINFONI observations of a unique, gravitationally lensed galaxy (from the KICC 2020 annual report).

One of the main challenges faced in modern astrophysics is explaining the major phase transition that transformed the very early neutral Universe into an (almost) completely ionised state. This process, referred to as reionization, was set in motion with the formation of the first stars and galaxies, around 150 million years after the Big Bang (around redshift 20), and was completed in under a billion years later (redshift around 6). (Read more about galaxies and their place in a timeline of the Universe here.) The current hypothesis is that low-mass galaxies were the main culprit behind Reionisation: at that point in time, pristine gas – largely composed of hydrogen and helium, not yet enriched with heavier elements like carbon and oxygen – abundantly flowed into these galaxies and was rapidly converted into stars. Therefore, a significant fraction of the resulting stellar populations were likely metal-poor, massive (O- and B-type) stars, whose spectra are much ‘harder’ than stars like the Sun: they produce copious amounts of photons with sufficiently high energy to ionise neutral atoms.

The challenge of confirming this hypothesis lies in the fact that as a result of their low stellar mass, these galaxies are necessarily faint, and thus hard to detect. To constrain their contribution to reionisation, an accurate measurement of two quantities is required. Firstly, observations need to reveal their occurrence rate, which can, in principle, be established relatively easily by number counts in deep surveys – this will certainly be achieved by the much-anticipated James Webb Space Telescope (JWST). Secondly, however, there is the slightly more complex question of their average ionising ability: did these systems produce enough ionising photons, and critically, could these escape the galaxy? This measurement is hindered by the very fact the Universe was in a neutral state initially, since intervening neutral gas (which absorbs all ionising photons) precludes us from directly observing the process of reionization. (Moreover, Lyman-alpha – the principal emission line of hydrogen and hence one of the main ways to identify high-redshift galaxies via their spectra – is also absorbed.)

Characterising sources that are ‘leaking’ ionising radiation is therefore crucial to aid JWST in finding the sources responsible for reionisation. To this end, we obtained new VLT/X-shooter spectroscopic observations (PI: Renske Smit) of a unique gravitationally lensed galaxy at redshift 5 (top figure on the left). A massive foreground cluster (RCS 0224–0002) serendipitously magnifies the background galaxy, named RCS0224z5, by a factor of about 30; only two other known sources at such high redshift have a comparably high magnification. Through measurements of rest-frame energetic ultraviolet emission lines, we found the radiation field to be predomi- nantly stellar in origin. Further analysis of previous observations taken with VLT/SINFONI allowed a measurement of neon and oxygen emission lines, which revealed a high number of ionising photons relative to the gas density inside the galaxy, far exceeding what is typically seen in local galaxies and indeed similar to nearby ‘leakers’, a class of extremely rare galaxies from which a significant fraction of ionising photons escapes. Furthermore, these lines are an indicator for the enrichment of heavy elements, which shows a tentative departure from the Fundamental Metallicity Relation (bottom figure on the left). Finally, we detected a magnesium emission line that has been shown to correlate with the same class of leakers. Independently, we also indirectly inferred that RCS0224z5 has a significant Lyman-alpha (and thus likely also ionising photon) escape fraction. Being the most distant galaxy for which this magnesium line has been reported, this demonstrates for the first time its potential as an indirect tracer of ionising photon leakage at high redshift.

These results are published as Witstok, J., et al., 2021, MNRAS, 508, 1686.

The following is a short description of research started in my master’s degree research project, in collaboration with Dr Ewald Puchwein, Dr Girish Kulkarni, Dr Renske Smit, Prof Martin Haehnelt, and Lewis Weinberger.

The gas outside galaxies and their direct environments, collectively called the intergalactic medium, forms the bulk of visible matter in the Universe. Thus far, however, it has mostly eluded direct detection by even the most powerful telescopes (although the first studies are starting to uncover its faint glow). Our predictions for observing this signal (see figure on the right) with the MUSE spectrograph and similar spectrographs on the next generation of observatories are published as Witstok, J., et al., 2021, A&A, 650, A98.

Lyα surface brightness (SB) in simulated narrowband images in a simulation snapshot at z = 4.8. The SB shown is a combination of recombination emission (of all gas in the simulation) below the mirror limit (indicated on the colour bar as HM12), and collisional excitation of gas below half the critical self-shielding density. Panel a shows a region of 8 x 8 (cMpc/h)^2 (5.2 x 5.2 arcmin^2) on a pixel grid of 1024 × 1024. Panels b-d show Lyα narrowband images the size of 1 x 1 arcmin^2 consisting of 300 x 300 pixels (as the field of view of MUSE). The volume probed by one of these narrowband images at this redshift is 2.84 (cMpc/h)^3. The areas covered by these maps are indicated by the white squares in the overview panel a. Halos with halo mass greater than 10^9.5 solar masses are shown as circles, their size indicating their projected virial radius. The most massive halo in each panel is annotated. Figure 8 from Witstok et al. (2021a).