A Spatial Gap in the Sky Distribution of Fast Radio Burst Detections Coinciding with Galactic Plasma Overdensities

The CHIME telescope is a compact radio interferometer that consists of 1024 dual-polarization antennas operating in the 400–800 MHz range. These antennas are distributed across four half-cylindrical reflectors that collectively span 80 × 100 m. The CHIME/FRB back end receives up to 1024 streams of beamformed, total-intensity data with time and frequency resolutions of t_samp ≈ 0.983 ms and Δν_chan ≈ 24.4 kHz, respectively (CHIME/FRB Collaboration et al. 2018), and enacts a series of real-time pipelines to identify and preserve astrophysical signals for further analysis. The CHIME/FRB detection pipeline uses a cutoff of signal-to-noise ratio (S/N) (≥(S/N)_thresh = 8) for filtering FRB candidates (CHIME/FRB Collaboration et al. 2025).

The celestial position of each FRB is computed from total-intensity data using the procedure described by CHIME/FRB Collaboration et al. (2019). This procedure uses two pieces of information—the real-time metadata corresponding to the detecting beam, and an analytic model of the beam shape and FRB dynamic spectrum—to estimate a position and its statistical uncertainty. These localizations have typical uncertainties of . We caution that O(1%) of bright Catalog 2 FRBs may be incorrectly localized to the near sidelobes, which can cause a position offset of 05–2°. This localization error does not have a significant impact on our analysis of the detection gap, as the error magnitude is much smaller than the angular extent of the gap. However, this could impact the classification of H ii intersections described in Section 5.2. We encourage a follow-up of this work using more accurate localizations.

Statistically significant signals detected by the CHIME/FRB back end are classified as “Galactic,” “ambiguous,” or “extragalactic,” based on the source position and DM estimated by the real-time bonsai tree dedispersion algorithm (CHIME/FRB Collaboration et al. 2018). Single pulses from Galactic radio pulsars are identified through coincidences in both position and DM and for the purposes of Catalog 2 are excluded from subsequent analysis. Our signals of interest are those deemed extragalactic (i.e., FRBs), which are classified as such if their DMs exceed both of the predicted maximum values from the NE2001 (J. M. Cordes & T. J. W. Lazio 2002) and YMW16 (J. M. Yao et al. 2017) electron density models for their LOSs; these predictions are made assuming a distance of 25 kpc between the observer and the edge of the Milky Way disk. No aspects of scatter broadening are considered for the real-time classification of FRBs.

In order to characterize the effects of the WIM on burst morphology in Catalog 2, we use spectrotemporal measurements of FRB morphology estimated with fitburst (E. Fonseca et al. 2024). The fitburst modeling framework assumes that each pulsed feature undergoes cold-plasma dispersion and, in the absence of scatter broadening, has a Gaussian shape of intrinsic temporal width (σ). If relevant, the shape of the burst is instead assumed to be the pulse broadening function of a Gaussian profile (e.g., M. M. McKinnon 2014) that depends on σ and τ_sc.

All Catalog 2 fitburst measurements assume that τ_sc scales with electromagnetic frequency (ν) such that τ_sc ∝ ν⁻⁴. The Catalog 2 values for τ_sc are originally reported referenced to ν = 400.195 MHz (CHIME/FRB Collaboration et al. 2026); however, for the present study we scale τ_sc to be referenced at ν = 1 GHz using τ_sc,ν ∝ ν⁻⁴ (N. D. R. Bhat et al. 2004), as is typically done for related studies on radio pulsars and FRBs. Additionally, for detectability discussions, we reference τ_sc to CHIME frequencies whenever relevant—explicitly denoting it as τ_{sc, 600 MHz}. We used τ_{sc, 1 GHz} > 0.13 ms as a threshold for determining which FRBs have “well-measured” scatter broadening; this threshold value was obtained through comparisons of fitburst fits between total-intensity and microsecond-resolution baseband (voltage) data of the same detections (K. R. Sand et al. 2025).

We adopted the initial data quality flags outlined in the Catalog 2 methodology (CHIME/FRB Collaboration et al. 2026). In addition, we required that the event have successful fits from the header_localization and fitburst pipelines. We excluded FRBs detected in the sidelobes of the primary beam, since these are more complicated to localize accurately (H.-H. Lin et al. 2024). The above selection criteria resulted in 4429 FRBs. To obtain the full set of unique sight lines from Catalog 2, we retain only the burst with the highest S/N for each repeating source. The final sample utilized in this study consists of 3542 FRBs—comprising 3459 nonrepeaters and 83 unique sight lines from repeating FRBs. Of those, 1105 FRBs meet the criterion for well-measured τ_{sc, 1 GHz} described in Section 2.4.

While Cygnus X is composed of many distinct star-forming regions, studies of molecular line observations have argued that they are associated, and at a similar distance d_Cyg (N. Schneider et al. 2006). This association was later supported using parallax measurements with masers (K. L. J. Rygl et al. 2012) and with Gaia (S. R. Berlanas et al. 2019), finding distance measurements spanning a range 1.3 kpc < d_Cyg < 1.7 kpc. We adopt a value of d_Cyg = 1.5 kpc throughout our work (following the discussion of F. Comerón et al. 2020), but we emphasize that it does not strongly affect our analysis. At d_Cyg = 1.5 kpc, the ∼18° angular diameter of the spatial gap corresponds to ∼250 pc. Additionally, the electron temperature T_e is a function of Galactocentric radius R_gal, fit by C. Quireza et al. (2006) as T_e(K) = (5780 ± 350) + (287 ± 46)R_Gal. We adopt T_e = 8000 K throughout for the Cygnus X region, appropriate for R_Gal ≈ 8 kpc.

4.1.1. Emission Measure

From the Planck emission measure (EM) map (Planck Collaboration et al. 2016), which traces the free–free emission from ionized gas and has a resolution of pc, the maximum EM of the Cygnus X region is 9400 pc cm⁻⁶. On smaller angular scales, the Cygnus X region contains many thin filamentary structures, which are overdense and overpressured compared to the average properties of their surroundings (K. L. Emig et al. 2022). The filaments reach a maximum EM of 10⁵ pc cm⁻⁶, and electron densities of n_e≈ 10−400 cm⁻³, compared to their respective medians of 5200 pc cm⁻⁶ and n_e ≈ 35 cm⁻³. The density of the volume-filling ionized gas in the region is n_e ≈ 35 cm⁻³, which itself is far above the volume-averaged plasma density in the WIM of n_e ≈ 0.03 cm⁻³ (P. O. Mezger 1978; K. L. Emig et al. 2022).

The observables and are physically related quantities, where n_e(l) is the free electron density (in cm⁻³) at a distance l (in pc) along the LOS, and L is the total path length through the ionized medium (in pc). Following S. K. Ocker et al. (2024), we assume that a path length L with volume-average electron density n_e is filled with dense cloudlets with a volume filling factor f ≤ 1 and internal density n_e,i, such that n_e = fn_e,i. The other parameters are the variance of electron density within a cloudlet and that between cloudlets . Following this prescription, the DM and EM are related as

The prefactor is always ≤1, forming an upper limit on the DM.

Based on the above, a source seen through Cygnus X at the maximum EM ≈ 10⁴ pc cm⁻⁶ (from the Planck map) with path length L ≈ 500 pc could accrue DM ≲ 2200 pc cm⁻³. As will be described in Sections 4.1.3 and 4.1.4, this estimate is in line with the highest DMs of pulsars in the region and that of the background FRB 20210705 discovered at L band with the Five-hundred-meter Aperture Spherical radio Telescope (FAST). As discussed in Section 4.1.1, the peak EM in filaments reaches 10⁵ pc cm⁻⁶, although most filaments have thickness on the order of a few parsecs (K. L. Emig et al. 2022). A sight line could, in principle, accumulate a larger DM if it is seen directly through the long axis of a filament, but this is unlikely and not representative of an average LOS through the Cygnus X region.

4.1.2. Angular Broadening

Background quasars, if intrinsically compact enough, can show signs of angular broadening when observed through multifrequency very long baseline interferometry (VLBI), characterized by broadening θ ∝ ν^α, scaling with α ≈ −2. Early VLBI observations of background quasars show angular broadening with full width at half-maximum (FWHM) of θ_FWHM ≈ 10–200 mas referenced to 1 GHz within Galactic longitudes 60° ≲ l ≲ 80°, greater than the surroundings (A. L. Fey et al. 1989, 1991). Observations of Cygnus X-3 (not associated with but background to the Cygnus X region) show scatter broadening of 220–250 mas at 1 GHz (L. A. Molnar et al. 1995). Many additional VLBI studies of quasars through Cygnus X (e.g., K. M. Desai & A. L. Fey 2001; T. J. W. Lazio & A. L. Fey 2001; K. É. Gabányi et al. 2006) show angular broadening of a similar order, as well as extreme-scattering-event-like structures seen in the light curve of quasar B2005+403, implying ∼0.7 au plasma structures (T. A. Koryukova et al. 2023).

Angular broadening, combined with the known distance to Cygnus X, can be used to predict the scattering time delay. For a scatter-broadened image with θ_FWHM, the scattering timescale τ_sc, defined as the 1/e timescale for a decaying exponential, is related to θ_FWHM as (e.g., J. M. Cordes & S. Chatterjee 2019)

where for a scattering screen at distance d_scr from the observer and source at distance D. For an extragalactic source and a thin screen near the observer, d_scr ≪ (D − d_scr) and d_eff ≈ d_scr.

This equation is exact for an angular intensity distribution following a Gaussian and can differ by a factor of ≈0.6–2 depending on the exact form of the scatter-broadened image. Scattering timescales derived for background radio sources using Equation (5) with d_scr = d_Cyg are listed in Table 1.

4.1.3. Pulsars

There are >90 known pulsars within our detection gap (R. N. Manchester et al. 2005). With surveys from more sensitive telescopes, many newly discovered pulsars in this region have DMs greater than the full expected Milky Way contribution from Galactic-n_e models (J. L. Han et al. 2025), as the abundance of plasma within Cygnus X is not properly accounted for. Without a reliable n_e model along these sight lines, and with few independent distance measurements, it is difficult to know which pulsars are foreground or background to the star-forming complex. As the complex itself is likely the cause of the discrepancy, the sources with DMs far exceeding model predictions can be assumed to be within or behind Cygnus X.

The FAST telescope recently measured scattering of 17 pulsars within our detection gap, spanning a range of τ_{sc, 1 GHz} ≈ 1–300 ms, and DMs 200–950 pc cm⁻³ (W. C. Jing et al. 2025). Other pulsars with scattering measurements include PSR J2108+5001, discovered by the CHIME All-sky Multi-day Pulsar Stacking Search (C. Andrade et al. 2025) and located just outside of our detection gap boundary, and nine pulsars discovered with the Green Bank Telescope at 820 MHz (A. E. McEwen et al. 2024).

The maximum DM predicted by NE2001 and YMW16 at the center of the detection gap is ∼500 pc cm⁻³. Pulsars with DMs exceeding this value are almost certainly located within or behind Cygnus X. Their scattering timescales can therefore provide an estimate of the propagation effects experienced by background sources, including FRBs intersecting this region. A list of the literature scattering measurements for DM > 500 pc cm⁻³ pulsars across the Cygnus X region is compiled in Table 1 and in Figure 2 (including J2108+5001 (482 pc cm⁻³), which falls just below this threshold but as a CHIME-detected pulsar is relevant to the context of CHIME’s observing capabilities).

Zoom In Zoom Out Reset image size

4.1.4. FRBs at Higher Observing Frequencies

As described in Section 2.3, the CHIME/FRB pipeline uses Galactic DM predictions from NE2001 and YMW16 to classify events as Galactic or extragalactic. This raises a general concern that DM-based thresholds could cause nearby, low-DM FRBs to be missed, if the models overestimate the Galactic DM along the LOS. For example, FRB 20220319D (DM = 110.95 pc cm⁻³, distance = 50 Mpc), detected by the Deep Synoptic Array (DSA-110), exhibits a DM lower than the model predictions (V. Ravi et al. 2025). However, in the Cygnus X region, the situation is reversed and the Galactic-n_e models underestimate the DM (see pulsars in Section 4.1.3). As a result, pulsars with unusually high DMs are more likely to be misclassified as FRBs, not vice versa. Therefore, this classification bias cannot explain the absence of FRBs in the detection gap.

We acknowledge the possibility that the Cygnus X complex is so poorly accounted for in the models that they could underestimate the maximum DM by an order of magnitude, i.e., that the true value of along the LOS. At such high DMs, intrachannel dispersion smearing could partially emulate scatter broadening and reduce pulse brightness at low frequencies. However, we believe that DM smearing can be ruled out as the origin of the detection gap. Temporal broadening due to intrachannel dispersion smearing (J. M. Cordes & M. A. McLaughlin 2003) is given by

The CHIME/FRB detection pipeline searches to a maximum DM of 13000 pc cm⁻³ (CHIME/FRB Collaboration et al. 2018). In order to produce t_chan ≈ 8 ms at 600 MHz—comparable to the mean value of τ_sc measured in Catalog 2 (1 ms at 1 GHz; see Section 4.1.4)—a value of DM > 8000 pc cm⁻³ is required. This estimate is substantially higher than the DM inferred from the maximum EM in the Cygnus X region as discussed in Section 4.1.1.

Here we examine whether the FRB detection gap toward Cygnus X could be caused by free–free absorption. The free–free opacity is well approximated by (P. G. Mezger & A. P. Henderson 1967)

As discussed in Section 4.1, the Planck EM map for the Cygnus X region indicates that EM ≈ 9400 pc cm⁻⁶, while detailed studies of overdense and overpressured filaments show peak values of EM ≈ 10⁵ pc cm⁻⁶(K. L. Emig et al. 2022).

Inserting EM ≈ 10⁴ pc cm⁻⁶ for the Cygnus X region (S. K. Ocker et al. 2024) and T_e ≈ 8000 K, the opacity is τ_ff ≈ 0.02 at ν = 400 MHz, the bottom of our band. For the maximum measured EMs of EM ≈ 10⁵ pc cm⁻⁶, free–free absorption begins to matter, with τ_ff ≈ 0.2 at ν = 400 MHz, but it is still a small effect, and with τ_ff < 0.1 averaged across the band. The detection of background quasars at 350 MHz (e.g., T. J. W. Lazio & A. L. Fey 2001) and the prevalence of free–free emission at 148 MHz (e.g., K. L. Emig et al. 2022) also qualitatively suggest that free–free absorption is not dominating at low radio frequencies in this region. We therefore conclude that while free–free absorption could have a small effect on a background FRB, it is not the primary cause of the detection gap.

The spatial gap is toward Cygnus X (J. H. Piddington & H. C. Minnett 1952), a bright source of radio continuum at CHIME’s observing frequencies that could reduce sensitivity to a background burst. Using the Haslam 408 MHz radio sky temperature map (C. G. T. Haslam et al. 1982), scaled to the central frequency of the CHIME band with a scaling index of −2.6, the average sky temperature within the detection gap is 35 K. To capture the effect of higher temperatures present in smaller subregions within the area, we use the 99th percentile T_sky = 108 K.

The expression for (S/N)_b, the S/N of the broadened pulse,

follows directly from the standard radiometer equation as used in FRB and pulsar literature (e.g., J. M. Cordes & M. A. McLaughlin 2003; C. Patel et al. 2018), where S_i is the intrinsic flux density, n_p is the number of polarizations measured, Δν is the bandwidth, and β is the correction factor for digitization. Burst detectability for CHIME/FRB is governed by the intrinsic burst width W_i, intrinsic flux, DM, and τ_sc. The observed peak flux density for an FRB decreases as its broadened pulse width—W_b—increases, defined as (D. W. Gardenier et al. 2021)

where z₀ is the redshift and t_chan depends linearly on the DM as described by Equation (6) in Section 4.2.

To compute the expected number of bursts within the detection gap, we perform simulations using subsamples of bursts from Catalog 2, explicitly accounting for the effect of elevated sky temperature. To ensure approximately uniform exposure and sensitivity, we select all Catalog 2 bursts in the similar declination range of 30° < δ < 50°, resulting in 919 bursts. This sample serves as an empirical proxy for the intrinsic burst sample seen at this declination range in the absence of the influence of Cygnus X. From Section 3, the expected number of bursts originating behind the gap (i.e., before accounting for the elevated T_sky in the Cygnus X region) is 19, following a Poisson distribution. We generate 10⁶ simulated realizations as follows:

1. 1.  
   Generate a number of bursts originating behind the gap N_gap, drawn from a Poisson distribution P(λ = 19).
2. 2.  
   Randomly select N_gap bursts from the Catalog 2 subsample of 919 bursts, using their intrinsic burst width (W_i), sky temperature corresponding to their positions T_sky,i, and (S/N)_i as burst parameters.
3. 3.  
   Compute the S/N resulting from the changed sky temperature in the gap, as

   
4. 4.  
   For each realization, find the number of bursts detected in the gap, N_detected,gap, such that (S/N)_Cygnus ≥ 8.

From the simulations, we find that the mean number of bursts detected would be 〈N_detected,gap〉 = 6, when accounting for the above-average T_sky. We find that zero detections occur in fewer than 0.3% of the simulations (i.e., N_detected,gap = 0), indicating that increased sky temperature and Poisson fluctuations alone are insufficient to account for the observed gap toward Cygnus X.

As summarized in Section 4.1, the measurements to date of sources observed through the Cygnus X region show significant scattering (τ_{sc, 600 MHz} ∼ 10–1000 ms), based on the angular broadening of background quasars and Cygnus X-3, as well as the scattering of distant pulsars and the FRB detected by FAST. This already suggests that scattering can be deleterious for FRB detections at CHIME and is likely a significant contributor to the lack of CHIME/FRB detections in the gap. However, the magnitude of scattering, which depends on plasma fluctuations on tiny scales at ≪1 au, can drastically differ between sight lines. In this section, we quantify the effect of scattering on CHIME/FRB detections and in turn set a lower limit on the average scattering in the detection gap that would lead to zero detections.

We first searched the positional and fitburst data in Catalog 2 for a radial trend in measured τ_sc from the boundary of the Cygnus X region, observing no clear trend. In Catalog 1 injections, CHIME/FRB found a strong selection bias against temporally broad bursts (W_b > 10 ms, calculated at 600 MHz; M. Merryfield et al. 2023). The lack of a trend may be due to the narrow observable window in τ_sc, which is constrained by the lower limit set by the reliability cutoff (0.13 ms at 1 GHz; see Section 2.4) and by the upper limit (≥10 ms at 1 GHz) that corresponds to scattered pulses likely too broad to detect. Along with a limited number of detected FRBs, a statistically significant radial trend is difficult to establish without finer time resolution to probe a wider range in τ_sc. As a result, it is difficult to trace whether there is a gradual increase in scattering timescales for sight lines closer to the center of the detection gap.

From the simulations described in Section 4.4, we found that on an average six bursts are expected to be seen in the detection gap. These simulations account for the elevated sky temperature but not for scattering. Here we extend the simulation to assess the degree of scattering required to see zero bursts.

In the simulations outlined in Section 4.4, we introduce pulse broadening due to scattering using Equation (9). For each simulated realization, we compute the minimum τ_sc using Equation (8), such that the increased W_b and T_sky reduce the S/N of all FRBs within the detection gap to (S/N)_b < (S/N)_thresh = 8. The distribution of τ_{sc, 1 GHz} values obtained from these simulations is shown in Figure 4. We find that a mean value of τ_{sc, 1 GHz} = 5.59 ms or τ_{sc, 600 MHz} = 43.13 ms is sufficient to suppress all detections. Because all simulations adopt the 99th percentile T_sky measured within the void area, these are conservative lower limits on τ_sc. Using more typical, lower T_sky would require even larger τ_sc to achieve the same suppression in (S/N)_b. The maximum τ_sc predicted by Galactic-n_e models for the LOS through the center of the detection gap are = 0.1 ms and = 16 ms. This range of model-dependent maximum values of τ_sc is consistent with our inferred maximum derived from the simulations. Moreover, this consistency reinforces the idea that enhanced scattering in Cygnus X is likely to explain the detection gap, though our limit was obtained in a manner independent of either n_e model. However, we emphasize that these models are outdated, differ in their treatment of the Cygnus X region,²⁰ and lack accurate modeling of small-scale plasma fluctuations, so their predictions should be interpreted only as rough references.

Zoom In Zoom Out Reset image size

To statistically test whether highly scattered FRBs are underdetected in general owing to propagation through regions of elevated EM, we compared EMs at FRB positions with those along all sight lines visible to CHIME. Exposure weighting is applied using the CHIME/FRB beam exposure map (CHIME/FRB Collaboration et al. 2026), such that each sky position contributes proportionally to its observing time. Figure 6 shows this comparison using the Planck EM map (Planck Collaboration et al. 2016). The FRB distribution drops to zero for EM ≥ 2900 pc cm⁻⁶. This difference demonstrates a deficit of FRBs along high-EM sight lines that are associated with ionized regions in the Galaxy.

Zoom In Zoom Out Reset image size

We can ask what the above EM cutoff of 2900 pc cm⁻⁶ corresponds to in terms of τ_ff, DM, and τ_sc. Similar to the calculations in Section 4.3, τ_ff is negligible (≈0.003) in our band. Using Equation (4) and assuming L = 10 pc (a typical H II region path length; see Section 5.2), we estimate a DM contribution near the cutoff of ≈170 pc cm⁻³, which is too small to induce appreciable DM smearing. We estimate τ_sc using the empirical τ_sc−DM relation from J. Cordes et al. (2016), resulting in a rough lower bound of τ_{sc, 1 GHz} ≳ 0.5 ms, or τ_{sc, 600 MHz} ≳ 4 ms. We note that use of their relation while assuming DM yields τ_sc ∝ EM^2.25 for DM ≳100 pc cm⁻³; values of EM above the aforementioned cutoff yield values of τ_sc that rapidly rise to magnitudes that make FRBs undetectable at CHIME. We also note that pulsars behind H ii regions lie preferentially above the τ_sc−DM relation and that a future reassessment is therefore warranted (S. K. Ocker et al. 2024). While we caution that the logic mapping EM to τ_sc is indirect, the above results support the interpretation that FRBs are preferentially scatter-broadened beyond detectability along high-EM sight lines.

Similar to the treatment for pulsar sight lines intersecting H ii regions by S. K. Ocker et al. (2024), we utilize the Wide-field Infrared Survey Explorer (WISE) V2.4 catalog of H ii regions (L. D. Anderson et al. 2014). The WISE catalog contains over 8400 H ii regions and H ii region candidates, identified using the all-sky WISE satellite data. We match FRBs to WISE H ii regions by computing angular separations in Galactic coordinates and checking whether each FRB falls within twice the angular radius of any H ii region (<2θ_IR). We adopt this maximum impact parameter of 2θ_IR following the prescription of S. K. Ocker et al. (2024) to account for potential underestimation of the ionized gas extent by the infrared size, which may miss diffuse outer emission caused by radiation leakage (M. Luisi et al. 2019; J. Dey et al. 2024). In Appendix B, we report the angular separations (η) between H ii regions and intersecting FRB sight-line candidates, in units of θ_IR. Out of the 3542 FRBs in our sample, 36 are candidate sight lines that pass through the WISE H ii regions. We similarly examined the optical H ii region catalog HH14 (L. G. Hou & J. L. Han 2014) following the method described in S. K. Ocker et al. (2024). We do not include the results in our main analysis owing to the lack of angular radii in HH14; however, the results are consistent with those obtained using WISE.

Among candidate FRB sight lines that intersect H ii regions, as well as those that do not but lie near the Galactic plane (∣b∣ < 10°), only about one-third pass the τ_sc reliability cutoff described in Section 2.4. This sample of sight lines is not sufficiently large to enable a robust statistical comparison of τ_sc between these subsets. For completeness, we report the mean and median τ_{sc, 1 GHz} values: for intersecting sight lines, the mean is 1.65 ms and the median is 0.63 ms; for nonintersecting near-plane sight lines, the mean is 1.06 ms and the median is 0.51 ms. The intersecting sample exhibits slightly higher values consistent with expectations, but an Anderson−Darling test between the two populations yields a statistic of 1.78 and a p-value of 0.06, indicating no significance. A more robust assessment will require either a larger sample from future CHIME/FRB catalogs or the use of raw baseband data with higher time resolution to permit measurement of shorter scattering timescales. For interested readers, Table 2 in Appendix B lists the FRB sight lines that intersect H ii regions, along with their positions, DMs, τ_sc, and angular separations in units of θ_IR. These may be useful for further case studies of scattering in dense Galactic environments. Their dynamic spectra are presented in Figure B1 in Appendix B to illustrate observed pulse morphologies.

For future investigation, we point out two interesting candidates. The FRB 20190912D LOS passes through the Sharpless 171 (S171) H ii region surrounding the Berkeley 59 cluster of young stars. The FRB 20200416A LOS passes through the giant W4 complex enclosing the Heart Nebula, hosting numerous O/B-type stars and the IC 1805 open cluster of young stars. Interestingly, the fitburst-measured τ_sc for FRB 20200416A is smaller than the maximum predicted by NE2001 for an extragalactic source along this LOS. This could indicate an overestimate by the model, or geometric suppression of scattering if the dominant screen lies at a cosmological distance/closer to the observer. However, we note that NE2001 was not designed to predict scattering with high precision, especially for extragalactic sources and along complex or poorly constrained sight lines. As such, this comparison should be interpreted cautiously.

We also explore the possibility of a correlation or lack thereof between τ_sc and DM for FRBs. Figure 7 shows a τ_sc–DM plot for all FRB candidates in Catalog 2 and exhibits an overall trend that confirms previous findings—FRBs are systematically underscattered compared to expectations from the τ_sc–DM relation from Galactic pulsars (J. Cordes et al. 2016). The discrepancy arises because a large fraction of the FRB DM comes from the intergalactic medium, which contributes minimally to scattering (D. R. Lorimer et al. 2007; D. Thornton et al. 2013). However, this comparison is biased to reflect the statistics of detected events, as we likely miss highly scattered FRBs passing through H ii regions that therefore fall below the CHIME/FRB detection threshold. For example, Catalog 2 FRBs that intersect Galactic H ii regions do not show a strong correlation between DM and τ_sc in Figure 7 for FRBs—even when considering only the Galactic contribution to DM—despite the expectation that significant δn_e in H ii regions causes scattering. However, FRB 20210705 as detected by the FAST telescope (see discussion in Section 4.1.4) indeed follows the τ_sc−DM trend, indicating that its scattering is consistent with being induced by the Cygnus X region. While the statistics currently remain small, this difference raises a cautionary note for population studies: low-scattering, high-DM bursts may be overrepresented in Catalog 2. CHIME/FRB’s selection function for Catalog 1 has been studied by injecting synthetic FRBs (M. Merryfield et al. 2023). This effort confirmed that the CHIME/FRB back end is more sensitive to bursts with smaller broadened pulse widths. An updated injection study for Catalog 2 is currently underway. In order to make conclusions about the cosmological population of FRBs, a follow-up study incorporating the updated selection functions and host models is essential.

Zoom In Zoom Out Reset image size

As discussed in 4.1.2, a scattered source will also be angularly broadened, where larger time delays correspond to larger deflection angles. Assuming d_eff ≈ d_scr for an extragalactic source, the mapping between delay τ and observed angle θ for each scattered path through a thin screen is (e.g., M. A. Walker et al. 2004)

The CHIME/FRB Outriggers project (CHIME Collaboration et al. 2025) augments CHIME/FRB by providing milliarcsecond-scale localization using VLBI across three strategically placed stations in North America. By enabling precise FRB localization, the Outriggers facilitate host galaxy identification, studies of FRB environments, and their use as cosmological probes. Given CHIME/FRB Outriggers’ target resolution of mas (CHIME Collaboration et al. 2025), FRBs scattered by τ_{sc, 600 MHz} ≳ 10 ms at ∼kiloparsec distances will have an angular extent larger than the Outriggers’ resolution. Many of the bright bursts piercing through H ii regions, or one detected through Cygnus X (with τ_{sc, 600 MHz} ≳ 40 ms), would be easily resolved by CHIME/FRB Outriggers. We note that while CHIME/FRB is biased against finding bursts of these scattering times, particularly bright bursts are seen with sufficient scattering for this approach (see, e.g., K. Shin et al. 2024 for a bright burst scattered ≳1 s in the CHIME band).

Given the mapping between τ and θ, the interferometric visibility (V) of scattered FRBs (or pulsars) will show a characteristic shape as a function of V(τ, ν) through a scattering tail (O. Wucknitz 2012, 2014), which allows for a determination of screen distances and orientations. Conversely, for spatially unresolved sources, an upper limit on the angular broadening will place a lower limit on the screen distance. This will allow mapping of Galactic scattering structures to astrophysical sources and will cleanly distinguish between Galactic and extragalactic scattering screens.

Upcoming ultra−wide-band instruments such as the Canadian Hydrogen Observatory and Radio Transient Detector (CHORD; 300–1500 MHz; K. Vanderlinde et al. 2019) and DSA-2000 (0.7–2 GHz; G. Hallinan et al. 2019) will be particularly well suited to studies of scattering, offering improved sensitivity across a broad frequency range.

Table 2 compiles the observational parameters of CHIME-detected FRB sight lines passing through H II regions.

Figure B1 presents the dynamic spectra for FRBs with sight lines intersecting Galactic H II regions to demonstrate the observed pulse morphologies.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size