TIME: Ionized-Carbon Tomographic Mapping

• TIME is a line-intensity mapping survey that tomographically traces [C II] emission at z≈6–9 and multiple CO rotational lines at lower redshifts.
• It employs a dedicated millimeter-wave instrumentation platform featuring a grating-spectrometer array and 1920 TES bolometers to achieve deep integrations and precise foreground mitigation.
• Its analysis techniques, including clustering and shot-noise power spectra measurements, constrain star formation during the Epoch of Reionization and the evolution of cosmic molecular gas.

The Tomographic Ionized-carbon Mapping Experiment (TIME) is a line-intensity mapping (LIM) survey designed to tomographically map the aggregate emission of the [C II] 158 μm fine-structure line at redshifts $6 \lesssim z \lesssim 9$, corresponding to the Epoch of Reionization (EoR), as well as multiple CO rotational lines from galaxies at $0.5 \lesssim z \lesssim 2$. Utilizing spectroscopic mapping with a dedicated millimeter-wave instrumentation platform, TIME enables constraints on both reionization-era star formation and cosmic molecular gas by coherently measuring both clustering and shot-noise fluctuations in unresolved line emission. The experiment is optimized for sensitivity, spectral coverage, and foreground mitigation in the presence of strong CO interloper emission from lower-redshift galaxies, leveraging a modern framework of empirical galaxy evolution, intensity mapping theory, and high-performance cryogenic detector arrays.

1. Instrumentation and Survey Design

TIME employs the ALMA 12 m Prototype Antenna at Kitt Peak, using a grating-spectrometer array comprising 32 spatial pixels, attaining a field of view of $14'$ by $0.43'$ (180 beams × 1), with a beam FWHM $\theta_{\mathrm{FWHM}}{\simeq}0.43'$ at 237 GHz. Spectral coverage spans 183–326 GHz at $R \equiv \nu/\Delta\nu \approx 90$–120, corresponding to $\Delta\nu \approx 2$–3 GHz per channel and $\Delta z \lesssim 0.01$ at 250 GHz, enabling coverage for [C II] at $6 \lesssim z \lesssim 9$ and CO(J = 3–6) at $0 < z < 3$. The mapped area is $1.3^\circ \times 0.43^\circ$, with dedicated “line scan” strategies achieving deep integrations: 1000 hr for TIME, expandable to $>$3000 hr for future extensions. The focal plane is populated with 1920 silicon-nitride isolated transition-edge sensor (TES) bolometers, split into high-frequency (HF; 230–325 GHz) and low-frequency (LF; 183–230 GHz) arrays. The nominal Noise-Equivalent Intensity (NEI) is $\sim$5 MJy sr⁻¹ s$^{1/2}$, and on-sky channel noise levels reach NET$_{\rm sys} \approx 200$–300 $\mu$K$\cdot\sqrt{\rm s}$ per pixel (Sun et al., 2020, Butler et al., 2 Oct 2025).

2. Detector Architecture and Performance

TIME’s detector subsystems deploy gold-mesh absorbers on low-stress SiN$_x$ membranes, in thermal contact with Ti/Al bilayer TES islands ($T_c \approx$ 480–510 mK). The optical coupling is via profiled feedhorns and planar dispersive spectrometers (4×8 for LF, 4×12 for HF), with absorber webs sized to optimize broadband efficiency (LF: 520 $\mu$m, HF: 480 $\mu$m). Four SiN$_x$ support legs ($\sim$2 $\mu$m $\times$ 500 $\mu$m) act as thermal and mechanical isolators.

Critical metrics are:

Parameter	LF Band	HF Band (Old)	HF Band (Redesign)
$T_c$ (mK)	$480 \pm 10$	$510 \pm 15$	$\sim$500
$G$ (pW/K)	$20 \pm 5$	$30 \pm 10$	$30$ (target)
$\eta_{\rm opt}$ (%)	$30$–$40$	$\sim$20	$30$–$40$ (goal)
Yield	$\gtrsim 90\%$	$60$–$70\%$	$\geq 85\%$

HF channel performance was initially limited by wafer thickness and backshort optimization; redesigns involving optimized backshort distance (370 $\mu$m), improved Kapton™ cable design (series resistance reduced from 10 mΩ to $\sim$6.6 mΩ), and wet-etch oxide clearance are expected to deliver LF-comparable efficiencies and yield. Laboratory characterization utilizes hot-cold load optical efficiency tests, I–V curves for thermal parameters, and on-sky calibration with principal component analysis for atmospheric removal. The fast response time (τ $\sim$ 1–5 ms) enables rapid scanning and tomographic mapping without significant beam smearing (Butler et al., 2 Oct 2025).

3. Scientific Objectives and Modeling Framework

TIME targets two key cosmic epochs:

1. Epoch of Reionization ($6 \lesssim z \lesssim 9$): [C II] intensity maps provide a direct measure of the cosmic SFR density in the faint, unresolved galaxy population. The LIM approach enables measurement of power spectra (auto and cross), yielding parameter constraints on the UV-luminosity–halo-mass relation (parameterized by $\xi$), escape fraction ($f_{\rm esc}$) of ionizing photons, and the [C II]–SFR scaling relation. Marginalized 68% uncertainties with the current design are $\sigma_{C\,\textsc{ii}} = 0.44^{+0.24}_{-0.27}$ dex and $f_{\rm esc} = 0.14^{+0.23}_{-0.08}$.
2. Cosmic Noon ($0.5 \lesssim z \lesssim 2$): Simultaneous mapping of CO($J=3$–6) yields robust constraints on the molecular gas density $\rho_{\mathrm{H}_2}$ via observed CO–LIR relations, including $\alpha=1.28\pm0.04$ and $\beta=-0.90\pm0.50$. Cross-spectra CO($J$) × CO($J+1$) are forecast at high S/N (20–26). CO–galaxy cross-correlations provide independent estimates of $I_{\rm CO,gal}$, supporting interloper validation (Sun et al., 2020).

The core modeling equations for mean intensity, shot noise, and clustering power follow: $$\langle \bar I_\ell (z) \rangle = \frac{c}{4\pi} \frac{1}{\nu_{\rm obs} H(z)} \int dL_\ell\, \Phi(L_\ell,z)\,L_\ell,$$

$$P_{\ell\ell}(k,z) = \langle \bar I_\ell \rangle^2\,b_\ell^2\,P_m(k,z) + P^\text{shot}_{\ell\ell}(z),$$

where $P^\text{shot}_{\ell\ell}$ involves the luminosity function second moments.

4. Foreground Mitigation and Data Analysis Strategies

A principal challenge for [C II] LIM at $6 \lesssim z \lesssim 9$ is contamination by interloper CO emission from foreground galaxies ($z<3$). The strategy developed by Sun et al. (2016) implements a voxel-based masking algorithm leveraging empirical IR–$M_*$–$z$ relations from the COSMOS/UltraVISTA $K$-band catalog and CO–LIR conversions (Greve et al. 2014), explicitly fitting: $$\log_{10} L_{\rm IR}(M_*,z) = \sum_{p=0}^n \Bigl[ \sum_{q=0}^n A_{p,q} (\log_{10} M_*)^q \Bigr] z^p$$ with adopted scatter $\sigma_{\rm tot} \sim 0.5$ dex. Voxels containing galaxies exceeding a redshift-dependent CO(4–3) flux threshold are masked, corresponding to mass cuts $M_* \gtrsim 10^9 M_\odot$ at $z \lesssim 1$ and $M_* \gtrsim 10^{10.5} M_\odot$ at $z = 2$, typically removing $f_{\rm mask} \approx 8\%$ of voxels. This yields CO power suppression $>90\%$ at $k=0.1\,h\,\mathrm{Mpc}^{-1}$ and [C II]/CO power ratio $\gtrsim10$, with total volume loss $\lesssim8\%$ and negligible ($<10\%$) S/N impact on [C II]. Further refinements can exploit photometric redshift PDFs, direct stellar-mass catalogs, or cross-correlation of residual maps for hierarchical interloper removal (Sun et al., 2016).

5. Optical System Characterization and Beam Mapping

Precise far-field beam characterization is critical for instrument calibration, systematic error control, and absolute intensity determination. For TIME, a planar cable-driven parallel robot (CDPR) “beam mapper” enables automated, repeatable mapping of submillimeter beams over a $\sim$400 mm × 400 mm optical workspace. The system consists of a lightweight, adjustable frame compatible with various mirror envelope geometries (600–800 mm wide), four stepper-driven cable actuators, and a 40 g payload PCB carrying 13 chopped thermal sources. Real-time position accuracy is tracked with non-contact computer vision (OpenCV-based ArUco and ChArUco fiducials), achieving root-mean-square error (RMSE) $<$2.7 mm in-plane and $<$0.5 mm repeatability. Out-of-plane (z) errors were $\leq$5.7 mm, all well within the beam profile sampling requirements ($xy_\mathrm{req}=10$ mm, $z_\mathrm{req}=5.9$ mm). Compared to handheld mapping, the CDPR supports $>$10× faster data acquisition and sub-mm repeatability, with successful deployment on field telescopes for rapid diagnosis and calibration (Mayer et al., 12 Nov 2025).

6. Data Analysis, Cross-Correlation, and Future Prospects

Analysis pipelines compute three-dimensional intensity data cubes, extract auto- and cross-power spectra, and use external galaxy catalogs for masking and cross-validation. Notable science targets include:

• [C II] auto-power: S/N$\sim$5–6 (first-generation); up to $\sim$20–30 for future arrays.
• $\rho_{\mathrm{H}_2}(z)$ at $z\sim0.6$–1.6 measured to $\lesssim$30%.
• CO–[C I] cross-powers and CO–galaxy cross-powers, with rigorous quantification of bias and mean intensity.
• [C II]–LAE or [C II]–21 cm cross-correlations possible in next phases to directly probe reionization bubble statistics.

TIME is foundational for larger, more sensitive surveys (e.g., TIME-EXT, TIME-NG), which will deploy on larger-aperture telescopes and adopt advanced mm-wave spectrometers (e.g., SuperSpec, DESHIMA on-chip systems). These will realize %%%%89$1.3^\circ \times 0.43^\circ$90%%%% more pixels, $>3\times$ lower NEI, and order-of-magnitude larger survey areas, enabling high-significance detection of reionization-era structure and model differentiation (Sun et al., 2020).

7. Significance and Outlook

TIME introduces a scalable, robust platform for [C II] and CO tomographic intensity mapping, addressing both the technical challenges of detector design, foreground mitigation, and calibration, and the astrophysical aim of constraining star formation and molecular gas evolution across critical cosmic epochs. This integrated approach serves as a blueprint for future LIM experiments targeting the early universe and paves the way for cross-disciplinary synergy with galaxy surveys, CMB probes, and 21 cm experiments (Sun et al., 2016, Sun et al., 2020, Butler et al., 2 Oct 2025, Mayer et al., 12 Nov 2025).