The FAST HI 21-cm Absorption Blind Survey. III. OH Absorption Search in the 21-cm Absorber Sample and Continued HI Absorption Search

Technical Summary: The FAST HI 21-cm Absorption Blind Survey. III. OH Absorption Search in the 21-cm Absorber Sample and Continued HI Absorption Search

Problem and Motivation
The hydroxyl radical (OH) serves as a critical tracer for the physical and chemical conditions of the interstellar and circumgalactic medium, particularly in diffuse molecular gas where carbon monoxide (CO) may be optically thick or inefficient. While OH absorption against background radio continuum sources is the primary method for detecting OH in non-masing galaxies at cosmological distances, only six cosmological OH absorbers have been identified to date. This scarcity is attributed to insufficient telescope sensitivity, radio frequency interference (RFI), selection biases in previous surveys, and the intrinsic low abundance of OH. Furthermore, known OH absorbers have consistently been detected in H i 21-cm absorption, suggesting that H i-selected samples may provide an effective parent population for systematic OH searches. This study addresses the need for an unbiased, blind search for OH absorption to constrain the OH content in radio-selected H i absorbers and to test for correlations between OH and H i abundances without the selection biases inherent in targeted surveys.

Methodology
The authors utilized data from the Five-hundred-meter Aperture Spherical Telescope (FAST) to conduct a blind search for H i 21-cm absorption and a subsequent targeted search for OH 18-cm absorption.

• Data Sources: The study extended previous blind H i searches using 2024 and partial 2025 data from the Commensal Radio Astronomy FasT Survey (CRAFTS) and the FAST All Sky H i Survey (FASHI), covering 394.4 hours and 1622.1 deg². This was combined with data from the FAst neuTral HydrOgen intensity Map- ping ExpeRiment (FATHOMER).
• H i Detection: A blind search algorithm was applied to cross-matched NRAO VLA Sky Survey (NVSS) sources with flux densities >12 mJy. Candidates were identified via matched-filtering against Gaussian templates in XX and YY polarizations, requiring a combined velocity-integrated signal-to-noise ratio (S/N) > 5.5.
• Follow-up and Characterization: Seven new H i absorption systems were confirmed through follow-up ON-OFF tracking observations using the FAST L-band Array of 19 feed horns (FLAN). Physical properties (column density, spin temperature, covering factor) were derived by modeling absorption profiles with Gaussian functions and integrating optical depth.
• OH Search: For 19 H i absorption systems where the redshifted OH 18-cm lines (1612, 1665, 1667, and 1720 MHz) fell within the FAST L-band (1.0–1.5 GHz), the authors searched for OH absorption.
• Statistical Analysis: To account for the high fraction of non-detections (upper limits), the study employed survival analysis using Bayesian censored regression (linmix) to investigate correlations between H i column density ($N_{\text{H I}}$) and OH column density ($N_{\text{OH}}$), as well as the evolution of the $N_{\text{OH}}/N_{\text{H I}}$ ratio with redshift. Spectral stacking was also performed to derive stringent upper limits on OH abundance for associated, intervening, and combined samples.

Key Contributions and Results

1. Expansion of the H i Catalog: The survey identified four new H i absorption systems and re-detected three known ones, bringing the total FAST blind H i absorption catalog to 41 systems (18 associated, 12 intervening, and 11 undetermined).
2. First Blind Detection of OH Absorption: The survey successfully re-detected OH 1612 MHz absorption and OH 1720 MHz emission in the known system PKS 1413+135. This marks the first time OH absorption has been detected in a blind survey, validating the strategy of using H i-selected samples for OH searches. No new OH absorbers were identified in the remaining 18 systems.
3. Constraints on OH-H i Correlation: Applying survival analysis to the sample, the study found no statistically significant evidence for a correlation between $N_{\text{OH}}$ and $N_{\text{H I}}$ for either associated or intervening absorbers. Similarly, no significant redshift evolution was found for the $N_{\text{OH}}/N_{\text{H I}}$ ratio. The authors note that previous studies reporting such correlations may have been influenced by selection biases toward dust-rich, molecule-rich systems.
4. Stringent Abundance Upper Limits: Spectral stacking of the non-detections yielded the tightest constraints to date on the OH content in radio-selected H i absorbers. The 3$\sigma$ upper limits on the OH column density are $4.93 \times 10^{12}$, $1.64 \times 10^{12}$, and $1.72 \times 10^{12} \, T_{\text{ex}}/c_{f,\text{OH}} \, \text{cm}^{-2} \text{K}^{-1}$ for associated, intervening, and combined samples, respectively. Assuming $T_{\text{ex}}=10$ K and $T_{\text{s}}=100$ K, these correspond to abundance ratios of $[\text{OH}]/[\text{H I}] < 1.66 \times 10^{-8}$, $< 1.42 \times 10^{-8}$, and $< 0.90 \times 10^{-8}$. These limits are comparable to or lower than those in diffuse Galactic sightlines and significantly lower than those in known extragalactic OH absorbers.

Significance
The paper establishes that known H i 21-cm absorbers provide an effective parent sample for systematic OH absorption searches, a strategy that has now been validated by the first blind detection of OH absorption. The results demonstrate that while H i absorption is a necessary condition for OH detection, it is not sufficient, implying that the majority of H i absorbers are molecule-poor or that OH is confined to small, low-filling-factor clumps. The study highlights the importance of unbiased surveys in avoiding the selection biases that have characterized previous targeted searches, which often favored dust-rich environments. The stringent upper limits set by this work provide a new benchmark for future studies of molecular gas in H i-selected systems and underscore the necessity of larger, unbiased samples to definitively constrain the physical drivers of the OH–H i connection and the stability of fundamental physical constants. The authors conclude that future wide-bandwidth surveys with next-generation facilities (e.g., FAST Core Array, SKA) are required to expand these samples and achieve higher statistical precision.