CHIME/Slow overview and pilot survey: A new backend to search for second-duration radio transients with the CHIME telescope

We present an overview of CHIME/Slow, a real-time transient search backend under development to search for second-duration radio transients using the CHIME telescope, and results obtained from a pilot survey carried out using the prototype version of the search pipeline. The prototype CHIME/Slow pipeline was tested on archival data obtained in December 2022, January 2023 and February 2023 with a total on-sky time of 17 days with an instantaneous Field of View (FoV) of $\sim$13 deg$^2$ . In this pilot survey, we detected nine bursts, one from a new non-repeating source and eight from the known hyperactive repeating source FRB 20220912A. Out of these nine bursts, two bursts from the repeater were not detected by CHIME/FRB, while the non-repeater was detected in the side-lobe of a beam in the CHIME/FRB exhibiting shorter pulse width and narrower bandwidth compared to the CHIME/Slow detection. Here we report properties of the bursts, discuss the sensitivity and completeness of the current version of the CHIME/Slow pipeline, and outline future development to improve its performance. Finally, based on these results, we report the all-sky rate (95% credible region) of radio transients with pulse widths between 16 ms to 5 s, fluence above 5 Jy ms and observing frequency of 600 MHz to be between 184 and 4556 bursts sky$^{-1}$ day$^{-1}$.

💡 Research Summary

The paper introduces CHIME/Slow, a new real‑time backend designed to search for radio transients with durations between 16 ms and 5 s using the Canadian Hydrogen Intensity Mapping Experiment (CHIME) telescope. Existing CHIME/FRB pipelines are optimized for sub‑100 ms bursts and therefore miss longer, second‑scale events. CHIME/Slow addresses this gap by ingesting the 3‑bit Huffman‑compressed intensity stream from CHIME/FRB, down‑sampling it to three time resolutions (16 ms, 128 ms, 512 ms), and performing independent single‑pulse searches on each stream with the PRESTO suite.

Key technical steps include: (1) initial RFI channel masking based on kurtosis, skewness, and standard deviation; (2) further narrow‑band, periodic, and impulsive RFI removal using PRESTO’s rfifind with parameters tuned via injection tests; (3) dedispersion over DM = 50–3000 pc cm⁻³ (step sizes of 2 pc cm⁻³ for 16 ms data and 10 pc cm⁻³ for coarser data); (4) single‑pulse searches with boxcar widths up to 0.13 s, 0.6 s, and 6 s for the three resolutions, adopting a detection threshold of S/N ≥ 10; (5) clustering of candidate events in DM–time space using HDBSCAN to eliminate sparsely clustered RFI; and (6) classification of the brightest event in each cluster with the FETCH convolutional‑neural‑network classifier, retaining events with astrophysical probability > 0.7.

A pilot survey was conducted on archival CHAMPSS data collected in Dec 2022–Feb 2023, amounting to 17 days of on‑sky time and an instantaneous field of view of ~13 deg² (equivalent to a full‑sky exposure of one day). The pipeline detected nine bursts: eight from the known hyper‑active repeater FRB 20220912A and one from a previously unknown non‑repeating source (temporarily labeled FRB 20230204C). Notably, two of the repeater bursts were missed by the standard CHIME/FRB pipeline, illustrating the added sensitivity of CHIME/Slow to longer‑duration events. The non‑repeater was seen in a side‑lobe of a CHIME/FRB beam with a narrow bandwidth and short pulse width, whereas CHIME/Slow measured a broader bandwidth and a longer pulse, highlighting differences in RFI handling and time‑resolution between the two pipelines.

Sensitivity and completeness were quantified via extensive injection of synthetic bursts spanning a range of fluences (3–30 Jy ms), widths (16 ms–5 s), and DMs. Recovery rates reach ~80 % for fluences ≥ 10 Jy ms in the 128 ms and 512 ms streams, but drop below 50 % for fluences ≤ 5 Jy ms in the 16 ms stream. Using these completeness curves, the authors performed a Bayesian estimate of the all‑sky rate of second‑duration transients (pulse widths 16 ms–5 s, fluence > 5 Jy ms at 600 MHz). The 95 % credible interval is 184–4556 bursts per sky per day, substantially lower than the FRB rate but indicating that such events are not vanishingly rare.

Future work outlined includes: (i) retraining FETCH on a CHIME‑specific labeled dataset to improve astrophysical vs. RFI discrimination; (ii) optimizing RFI masking parameters and DM step choices to boost sensitivity; (iii) transitioning the pipeline from offline to real‑time operation using GPU/FPGA acceleration; (iv) expanding the DM and time‑width search space; and (v) integrating CHIME/Slow alerts with the existing CHIME/FRB system and other facilities (ASKAP, LOFAR, etc.) for multi‑wavelength follow‑up.

In summary, CHIME/Slow successfully demonstrates a viable approach to systematically search the CHIME sky for second‑scale radio transients, recovers events missed by the standard FRB pipeline, provides the first quantitative estimate of their sky rate, and sets a clear roadmap for enhancements that will enable a new window on the transient radio universe.