Through Thick and Thin: The Cosmic Evolution of Disk Scale Height

TL;DR

This study uses JWST/COSMOS-Web to measure global disk scale heights $h_0$ for $M_* \ge 10^{10}\,M_\odot$ edge-on galaxies across $0<z<3.5$, implementing a fixed rest-frame wavelength of $1\,\mu$m to minimize dust effects and a novel bias correction for projection. By fitting 2D bulge+disk models with a sech$^2$ vertical profile and constructing hybrid PSFs, the authors derive robust $h_0$ and $R_d$ values, and compare with a 1D approach and an exponential vertical profile to ensure consistency. After correcting for wavelength-dependent extinction and inclination biases, they find $h_{0,1\mu m}$ grows from $\sim0.56$ kpc at $z\sim3.25$ to $\sim0.84$ kpc at $z\sim1.25$, then declines to $\sim0.67$ kpc by $z\sim0.25$, while the inverse scale-height ratio $R_d/h_0$ remains roughly constant at $\sim2.7$ for $z>1.5$ and rises to $\sim4.0$ by $z\sim0$. The results imply that high-redshift disks were intermediate-thickness, dynamically hot systems that thickened through decreases in surface mass density and violent instabilities, followed by thin-disk growth dominating at $z\lesssim1$, leading to present-day thick–thin disk configurations. A simplified evolutionary model relating $h_0$ to the stellar mass surface density $\Sigma_{*,\rm disk}$ reproduces the observed trends and yields a rising thin-disk fraction toward the present day, with a peak transition around $z\approx1$. Overall, the work argues for an evolutionary link from early dense, hot disks to modern thick-thin disks, while disfavoring minor mergers as the primary driver of thick-disk formation.

Abstract

To investigate the formation and evolution of vertical structures in disk galaxies, we measure global $\operatorname{sech}^2$ scale heights, averaging thin and thick components when present, for 2631 edge-on disk galaxies with $M_*>10^{10} M_\odot$ at $0< z < 3.5$ from the JWST COSMOS-Web survey. We show that dust extinction systematically overestimates scale heights at shorter rest-frame wavelengths, and therefore adopt a fixed rest-frame wavelength of 1 $μ$m. After further correcting for projection-induced bias using a new accurate method, we find that the median disk scale height increases from $0.56\pm0.03$ kpc at $z=3.25$ to $0.84\pm0.04$ kpc at $z=1.25$, and subsequently decreases to $0.67\pm0.06$ kpc at $z=0.25$. The disk length-to-height ratio remains constant at $2.7\pm0.2$ for $z>1.5$, but rises to $4.0\pm0.4$ at $z=0.25$. These results imply that the high-redshift progenitors of present-day thick disks were of intermediate thickness, neither thin nor thick, yet dynamically hot and dense. The observed radial variation of scale height is consistent with the artificial flaring expected from observational effects, disfavoring minor mergers as the primary mechanism of disk thickening. Instead, we suggest that the high-redshift intermediate-thickness disks were single-component systems that increased their vertical scale height through decreasing surface mass density and/or violent gravitational instabilities, eventually producing thick disks. Thin-disk growth begins at $z\approx2$ and dominates at $z\lesssim1$, yielding a vertically more compact system with decreasing scale heights from $z\approx1$ to $0$. The inferred thin-disk mass fraction increases from $0.1\pm0.03$ at $z=1$ to $0.6\pm0.1$ at $z=0$. Together, these findings reveal a continuous evolutionary link between high-redshift single-component disks and present-day thick thin disk systems.

Figures (22)

• Figure 1: Observed rest-frame wavelengths of the NIRCam filters used in COSMOS-Web as a function of redshift. The shaded region marks the bandwidth of the filter. The upper redshift limit is chosen to ensure that the pivot wavelength of the F444W filter probes rest-frame wavelengths of at least 1 $\micron$. We obtain the scale heights at a fixed rest-frame 1 by interpolating the measurements across the available filters (see Section \ref{['sect:rest']}).
• Figure 2: Diagram of $M_r-M_J$ versus $M_{\rm NUV}-M_r$ for our edge-on disk galaxy sample (blue dots). Contours indicate regions enclosing a given fraction of galaxies in the parent sample within each redshift range. Quiescent galaxies, which mostly lie above the dividing line proposed by Ilbert2013, are excluded. Median stellar mass and scatter for our sample for each redshift bin are shown at the bottom.
• Figure 3: Comparison of the (top) ePSF and (bottom) our hybrid PSF for each JWST NIRCam filter of COSMOS-Web. The ePSF is constructed using PSFEX, while the hybrid PSF is constructed by replacing the low-$S/N$ pixels in ePSF with the values from simulated theoretical PSFs using a smooth top-hat function. These hybrid PSFs are adopted for all model fittings throughout this work.
• Figure 4: Example of the Sérsic-sech$^2$ model fitting. The top row shows a galaxy (ID = 364) at $z=0.54$, with the left and right three panels displaying the results for the F444W and F115W filters, respectively. The bottom row shows a galaxy (ID = 486) at $z=2.45$, with results for the F277W and F115W filters. A scale bar is shown at the bottom-left corner. The derived sech$^2$ vertical scale height ($h_0$) is indicated in each case.
• Figure 5: Correlation between the scale heights derived from the sech$^2$ function ($h_0$) and the exponential function ($h_{0.\rm exp}$). The two measurements are tightly correlated, with $h_0$ being on average 1.37 times as large as $h_{0,\rm exp}$.