Spectroscopic redshifts of selected flat-spectrum radio sources I

Abstract: We present the first results of a spectroscopic campaign carried out as part of the Search for Milli-Lenses (SMILE) program, which aims to constrain the prevalence of gravitational lens systems on milli-arcsecond angular scales (milli-lenses) using high-resolution Very Long Baseline Interferometry (VLBI) imaging. The SMILE parent sample contains ~ 5000 radio-loud active galaxies, selected as a flux-limited, complete subsample of CLASS (The Cosmic Lens All-Sky Survey) sources. We compiled redshift information for the full sample from multiple literature and catalog sources and found that 491 sources have no available redshift estimate, either spectroscopic or photometric. A further 948 sources have only photometric redshifts, many of which show substantial discrepancies between catalogs. Reliable redshifts are essential for VLBI radio-source studies because they convert angular measurements into physical linear scales, enable estimates of intrinsic luminosities and jet kinematics, and allow robust cosmological and population studies. To address this key limitation for lensing and population studies, we initiated a dedicated spectroscopic campaign to secure reliable redshifts for as many targets as possible. This paper focuses on the brightest sources in the SMILE sample. We report newly determined spectroscopic redshifts for 6 targets out of 15 observed with the Skinakas 1.3 m telescope.

• The paper presents a focused spectroscopic campaign that measured precise redshifts for flat-spectrum radio AGN, bolstering milli-lens searches.
• It employs low-resolution spectroscopy and Gaussian line modeling to achieve redshift uncertainties as low as 2×10⁻⁵, resolving discrepancies in photometric catalogs.
• Results improve lensing cross-section calculations and AGN physical characterizations, directly impacting dark matter substructure studies.

Spectroscopic Redshifts of Selected Flat-Spectrum Radio Sources: Methodology and Implications for Milli-Lens Searches

Introduction and Scientific Context

The paper addresses a fundamental bottleneck in high-resolution gravitational lensing surveys: the scarcity of robust spectroscopic redshifts for large samples of flat-spectrum radio-loud AGN used in milli-lens searches. The SMILE (Search for Milli-Lenses) project assembles $\sim$5000 sources as a complete flux-limited subset of the CLASS catalog for systematic VLBI studies of milliarcsecond-scale image separations. This angular regime is essential for constraining lensing by supermassive ($10^6$ to $10^9\,M_\odot$) compact masses and the abundance of subgalactic dark-matter structures predicted by $\Lambda$CDM [Zavala & Frenk 2019].

Robust redshifts are imperative for several reasons: lensing cross-sections require precise source distances for meaningful statistical constraints on dark-matter models; the conversion of measured angular separations into intrinsic linear scales is critical for rejecting false positives such as CSOs, which contaminate milli-lens samples but have projected physical sizes generally $<$1 kpc [Kiehlmann et al. 2024]; reliable luminosity and kinematic estimates for radio jets and lens modeling depend on accurate redshift determinations. Despite leveraging all major spectroscopic and photometric redshift catalogs, nearly 1500 sources in the sample lacked spectroscopically confirmed redshifts or had only highly uncertain photometric estimates. The paper reports a focused campaign to secure $z_{\rm spec}$ for the optically brighter portion of these radio sources.

Methodology: Spectroscopic Follow-Up and Data Analysis

Candidate selection was based on the absence of prior $z_{\rm spec}$ and $r$-band or Gaia $G_{\rm RP}$ magnitudes brighter than 19. Observations were conducted at the Skinakas Observatory 1.3 m telescope using a low-resolution spectrograph yielding $R\sim530$ and coverage of the key rest-frame optical emission lines typically seen in AGN host galaxies and quasars. SNR constraints were carefully documented, and a custom reduction pipeline implemented rigorous flat-fielding, cosmic ray rejection, wavelength calibration, and continuum normalization.

Line identification and redshift measurement were performed via Gaussian modeling of emission lines; blended features were fit by simultaneously modeling up to three Gaussian components. Line-based $10^6$0 estimates were combined using a weighted mean. For the majority of sources, redshifts were determined from strong lines such as H$10^6$1, [O III] $10^6$25007, [O I], [N II], [S II], and, when applicable, prominent UV lines at higher redshifts. Redshift uncertainties robustly propagate centroid, calibration, and fitting errors.

Empirical Results: Spectroscopic Performance and Catalog Comparison

New spectroscopic redshifts were determined for 6 of 15 observed AGN lacking prior robust measurements. Reported SNRs per pixel for these spectra range from 3 to 26, with redshift errors as low as $10^6$3. Measured redshifts are as follows:

• GB6 J010341+423925: $10^6$4
• GB6 J184835+213156: $10^6$5
• GB6 J195141+480145: $10^6$6
• GB6 J201414+063439: $10^6$7
• GB6 J203142+162147: $10^6$8
• GB6 J222252+144119: $10^6$9

For these sources, the spectroscopic $10^9\,M_\odot$0 typically resolves discrepancies among ($10^9\,M_\odot$1) catalog-based photometric redshifts, directly impacting physical size and lensing cross-section calculations for the SMILE sample. Notably, line diagnostics for several sources permitted highly robust $10^9\,M_\odot$2 determinations; e.g., for GB6 J195141+480145, the measured $10^9\,M_\odot$3 is an order of magnitude more precise than typical photometric methods.

Additional independent redshift confirmations were obtained for sources whose $10^9\,M_\odot$4 appeared in external catalogs after the campaign. Measured values were in excellent agreement (e.g., $10^9\,M_\odot$5 vs. catalog $10^9\,M_\odot$6).

For four sources, no emission lines were detected, consistent with expectations for BL Lacs or featureless AGN, or with SNR limitations. The analysis uncovered a likely foreground star associated with one radio position, underlining the need for spectroscopic vetting to eliminate contaminants.

Practical and Theoretical Implications

The spectroscopic redshifts secured in this campaign materially improve the reliability of SMILE's core science goals—particularly, robust separation of milli-lens candidates from contaminant populations and accurate calculation of lensing statistics for dark matter substructure studies. The comprehensive effort to compile, cross-match, and supplement redshift information exemplifies a necessity for large-sample VLBI projects: even in an era of extensive machine-learning-based cataloging, direct spectroscopic confirmation remains the gold standard for high-precision extragalactic science, especially when photometric values are inconsistent or missing.

These results underscore both the promise and the limitations of employing even modest-aperture telescopes in assembling comprehensive $10^9\,M_\odot$7 datasets for large radio AGN samples. The detection yield for emission-line AGN at $10^9\,M_\odot$8 observed during high jet continuum states was $10^9\,M_\odot$9 (6/15), indicating the remaining fraction may require larger apertures, improved SNR, or follow-up in different continuum states.

Future Directions

The program targets increased spectroscopic completeness for the full SMILE sample. Fainter objects or sources in persistently featureless states will require observations with larger telescopes and deeper exposures. As the SMILE program progresses, systematically improving redshift completeness will translate to more stringent upper limits (or detections) for the milli-lens population, crucial to testing $\Lambda$0CDM small-scale structure predictions [Loudas et al. 2022]. Cross-validation of future spectroscopic campaigns, improved automated vetting for contaminants (e.g., foreground stars), and integration with astrometric databases (e.g., Gaia) will further mitigate systematic uncertainties in lensing statistics and population studies.

Conclusion

This spectroscopic campaign substantially increases the robustness of redshift assignments for a key subset of the SMILE VLBI milli-lens sample (2607.04959). The resultant $\Lambda$1 values significantly outperform public photometric estimates in terms of precision and reliability, directly enabling improved physical characterization of radio AGN and accurate classification of milli-lens candidates. Systematic spectroscopic follow-up remains indispensable for cosmological lensing programs, even when advanced machine-learning and high-volume cataloging are available. The continued expansion of this redshift catalog will enhance the scientific return of ongoing and future milli-lens studies, informing both practical strategies for survey design and theoretical challenges in substructure lensing.