FRB 20180916B: Periodic Repeating FRB

• FRB 20180916B is a repeating fast radio burst localized in a spiral galaxy, exhibiting a stable 16.3-day periodic cycle and frequency-dependent active window.
• High-resolution localization and multiwavelength analysis have provided detailed constraints on its local environment and challenged simple young magnetar models.
• Multi-frequency observations reveal narrowband, structured bursts with evolving rotation measure, offering key insights into intrinsic burst physics and propagation effects.

FRB 20180916B, also designated FRB 180916.J0158+65, is a nearby repeating fast radio burst localized to the spiral galaxy SDSS J015800.28+654253.0 at redshift $z=0.0337$ and luminosity distance about $149$–$150$ Mpc. It is distinguished by a periodic activity cycle measured near $16.3$ days, a chromatic active window that shifts with radio frequency, and an unusually extensive set of radio, polarimetric, environmental, and multiwavelength constraints. These properties place it among the best-characterized repeaters and make it a central case for separating intrinsic burst physics from propagation- or geometry-induced modulation (Tendulkar et al., 2020, Espinoza-Dupouy et al., 7 Jul 2025).

1. Host galaxy, localization, and local environment

FRB 20180916B is localized with milliarcsecond precision in the massive spiral galaxy SDSS J015800.28+654253.0. HST/WFC3 imaging and GTC/MEGARA integral-field spectroscopy resolved its environment on $30$–$60$ pc scales and showed that the FRB lies in a spiral-arm star-forming complex, but not at the brightest compact knot within that complex. The radio position is offset by $0.355'' \approx 250$ pc from the brightest F110W pixel of the nearest young stellar clump, whose characteristic size is quoted as $\sim 380$ pc (full width at half maximum). The surrounding “V”-shaped structure is kinematically part of the rotating host disk rather than a detached satellite system (Tendulkar et al., 2020).

The burst site itself is comparatively faint in nebular emission. No point source is detected at the FRB position, and the H$\alpha$ limit implies $$L_{\mathrm{H\alpha}} \lesssim 8\times10^{36}\,\mathrm{erg\,s^{-1}}$$, quoted in the abstract as $149$0. The corresponding local star-formation-rate limit is $149$1. The same H$149$2 non-detection constrains any ionizing stellar companion: a single main-sequence star hotter than O6V is ruled out, while late-O, B, and Be-type companions remain allowed. The local metallicity is about half solar, with $149$3, and the host disk is nearly face-on with stellar inclination $149$4 (Tendulkar et al., 2020).

These environmental measurements disfavor the simplest picture of a very young isolated magnetar still embedded at its natal star-forming core. If the source originated in the nearest young clump, the observed projected offset implies a travel time of about $149$5 to $149$6 for typical neutron-star system velocities. That range is much older than the $149$7 active lifetime often invoked for newborn magnetars, but is compatible with high-mass X-ray binary or $149$8-ray binary timescales (Tendulkar et al., 2020).

2. Periodic activity and the chromatic activity window

The defining large-scale temporal property of FRB 20180916B is a periodic activity cycle of $149$9 days, with active and inactive windows rather than strictly periodic individual bursts. In the phase convention adopted by several studies, observations are folded relative to MJD $150$0, and the burst population clusters around mid-cycle. In a unified chromatic analysis based on 214 observing epochs from 2018 September 16 to 2021 December 1, the peak of the folded activity distribution at the reference frequency of 600 MHz is $150$1, and the active window is described by its frequency-dependent centroid $150$2 and FWHM $150$3 (Espinoza-Dupouy et al., 7 Jul 2025).

A central quantitative result is that the activity window shifts to earlier phase at higher radio frequency. The fitted centroid law is

$150$4

while the width is

$150$5

Because $150$6, higher-frequency activity occurs earlier in phase. By contrast, the width trend is weak and not statistically significant; the source is often summarized as broadening at lower frequencies or, equivalently, only weakly narrowing at higher frequencies (Espinoza-Dupouy et al., 7 Jul 2025).

Independent low- and high-frequency campaigns sharpen this chromatic picture. LOFAR monitoring at 110–188 MHz constrains the low-frequency activity window to $150$7 day with phase centre $150$8, substantially later than the CHIME band, while Effelsberg detections at $150$9–$16.3$0 GHz occurred within a $16.3$1-day window that began $16.3$2 days before the CHIME activity peak (Gopinath et al., 2023, 2207.13669). At 150 MHz, earlier LOFAR work described the delay relative to $16.3$3 MHz as about $16.3$4 days, and simultaneous CHIME/FRB bursts without LOFAR counterparts showed that the effect is not simply broadband emission extending coherently across the whole radio spectrum (Pleunis et al., 2020).

The long-term periodicity itself appears stable over the CHIME monitoring baseline. A CHIME/FRB study through 2021 December measured $16.3$5 d and found $16.3$6, consistent with zero; the quoted $16.3$7 upper limit on $16.3$8 is $16.3$9 (Sand et al., 2023). This stability is one of the main constraints on dynamical models.

3. Radio burst phenomenology across frequency

FRB 20180916B has been detected across an unusually broad radio range, from at least 110 MHz up to the $30$0–$30$1 GHz sub-band of Effelsberg observations. At the lowest frequencies, LOFAR detected bursts in the 110–188 MHz band, including emission visible down to 110 MHz. These bursts are broad, with widths of roughly $30$2–$30$3 ms, and their low-frequency broadening is well described by scattering: stacked LOFAR analysis gives $30$4 ms and $30$5. The same low-frequency data showed strong narrowband behavior, including five CHIME/FRB bursts during overlapping LOFAR windows with no LOFAR counterparts (Pastor-Marazuela et al., 2020, Gopinath et al., 2023).

At intermediate frequencies, coordinated uGMRT, CHIME/FRB, and GBT observations over 300–1000 MHz detected 12 bursts in 3 days: 4 with the uGMRT, 1 with CHIME/FRB, and 7 with the GBT. This campaign produced the first detection in the 800–1000 MHz range and identified $30$6 structure in one burst at 800 MHz, the lowest-frequency detection of such fine temporal structure for this source at the time. Several bursts showed downward-drifting sub-bursts, with representative drift rates of $30$7 MHz ms$30$8 at uGMRT and as large as $30$9 MHz ms$60$0 in the GBT band, reinforcing the empirical trend that drift-rate magnitude increases with observing frequency (Sand et al., 2021).

At still higher frequency, Effelsberg detected eight bursts in the $60$1–$60$2 GHz campaign, all in the $60$3–$60$4 GHz sub-band. These bursts were much narrower in time than their lower-frequency counterparts: most had measured widths only as upper limits of $60$5–$60$6 ms, with two resolved cases at 3.42 ms and 0.79 ms. Their spectral extents were hundreds of MHz, but still much smaller than the full receiver bandwidth. The same study reported that the burst rate scales with frequency as $60$7, although the authors cautioned that the high-frequency rate is likely underestimated because of time-resolution limitations and narrow burst bandwidths (2207.13669).

At microsecond resolution, EVN voltage data at 1.7 GHz showed the most extreme fine structure yet reported for the source. Four bursts were resolved with emission timescales spanning roughly three orders of magnitude, down to $60$8–$60$9, probably limited by scattering. The brightest burst contained $0.355'' \approx 250$0–$0.355'' \approx 250$1 fluctuations and a $0.355'' \approx 250$2 leading spike, while a different burst contained $0.355'' \approx 250$3–$0.355'' \approx 250$4 components detectable across the full 128 MHz band (Nimmo et al., 2020).

Taken together, the radio phenomenology is that of a narrowband, morphologically structured repeater whose instantaneous emission does not occupy the full accessible band even when the source is active over a broad frequency range.

4. Polarization, RM evolution, and the meaning of DM variability

Polarimetrically, FRB 20180916B is highly ordered. At 1.7 GHz, EVN observations showed that all four analyzed bursts were highly linearly polarized ($0.355'' \approx 250$5), with little or no circular polarization ($0.355'' \approx 250$6) and nearly constant polarization position angle within and between bursts. The measured RM for the anchor burst B4 was $0.355'' \approx 250$7, and the brightest microsecond-resolved structure showed only subtle few-degree PPA variations on $0.355'' \approx 250$8 timescales (Nimmo et al., 2020).

Later low-frequency monitoring revealed that the source’s Faraday screen is dynamic. CHIME/FRB baseband polarimetry of 44 bursts from 2018 December to 2021 December found that during 2021 April–December the RM changed by about $0.355'' \approx 250$9, with a fitted slope

$\sim 380$0

while the dispersion measure showed only $\sim 380$1. Because

$\sim 380$2

the combination of large RM evolution and small DM evolution was interpreted as evidence that the line-of-sight magnetic field in the Faraday-active medium changed more strongly than the total electron column density (Mckinven et al., 2022).

uGMRT monitoring over $\sim 380$3 days refined this picture further. In 116 Band-4 detections, of which 79 were polarization-calibrated, the early data continued the CHIME secular trend, but later observations indicated a transition to stochastic RM variations around a new approximately constant level near $\sim 380$4. The weighted mean linear polarization fraction for the calibrated bursts is $\sim 380$5. That study also noted a tentative sign flip in the host-frame RM contribution, depending on the adopted Galactic foreground subtraction (Bethapudi et al., 2024).

The relation between DM measurements and burst morphology is more subtle than a single number might suggest. The DM-power analysis of 11 bright uGMRT bursts obtained optimized effective DMs between $\sim 380$6 and $\sim 380$7 over a span of about 2 hr, with formal precisions as small as $\sim 380$8. However, the authors argued that this burst-to-burst spread is more likely produced by intrinsic frequency-dependent burst structure than by true line-of-sight electron-column changes (Lin et al., 2022). A plausible implication is that for FRB 20180916B, “the DM” is estimator-dependent once the analysis becomes sensitive to different temporal scales of complex burst morphology.

5. Multiwavelength searches and counterpart limits

FRB 20180916B is a particularly favorable target for counterpart searches because it is nearby and repeating, with a known activity cycle that permits scheduled simultaneous observations. The most sensitive millisecond-timescale optical study so far used Gemini-N/‘Alopeke during two CHIME-detected bursts, FRB 20201023 and FRB 20220908. No significant optical counterpart was detected. After extinction correction, the $\sim 380$9 optical fluence limits are

$\alpha$0

for FRB 20201023 and

$\alpha$1

for FRB 20220908, implying

$\alpha$2

with a best combined constraint of about $\alpha$3. Those limits rule out models in which FRB 20180916B has a Crab-like high optical-to-radio fluence ratio, and they also exclude some inverse-Compton scenarios, while remaining above the level predicted in synchrotron-maser outflow models (Kilpatrick et al., 2023).

A broader simultaneous panchromatic campaign covered radio, optical, X-ray, and $\alpha$4-ray bands across eight activity cycles between 2020 October and 2021 August. It detected 14 new bursts with the SRT at 336 MHz and 7 with the uGMRT at 400 MHz, but no secure non-radio counterpart. For the uGMRT burst uGMRT-03 observed simultaneously with TNG/SiFAP2, the non-detection implies

$\alpha$5

For the SRT burst SRT-P-02 observed simultaneously with Insight-HXMT, the 1–30 keV non-detection gives

$\alpha$6

depending on the assumed X-ray spectral model. AGILE, INTEGRAL, and Swift also found no significant counterpart during their covered intervals (Trudu et al., 2023).

These null results constrain only the brighter prompt counterpart parameter space. They do not exclude weaker magnetar-like multiwavelength behavior, but they strongly limit prompt optical emission on $\alpha$7 ms timescales and rule out a subset of models that predict relatively luminous broadband flashes simultaneous with the radio burst.

6. Physical interpretation and continuing controversies

The observational record of FRB 20180916B has been used to test models based on orbital motion, precession, ultra-long rotation, and mixed source-plus-environment frameworks. One clear empirical result is that simple companion-wind absorption models are strongly disfavored. Simultaneous Apertif and LOFAR work showed that the activity window is earlier and narrower at higher radio frequencies, whereas binary wind interaction models predict a narrower periodic activity window at lower frequencies. The same low-frequency detections down to 120 MHz show that low-frequency emission can escape the local medium, weakening models that rely on strong low-frequency absorption to create the periodicity (Pastor-Marazuela et al., 2020).

Later CHIME/FRB morphology work sharpened the same point from a different angle. Over three years, the source showed no phase-dependent DM or scattering trend despite strong activity modulation, and no DM or scattering response to the secular RM change. That paper concluded that the upper limits on phase-dependent changes in DM and scattering do not support models invoking a massive binary companion star as the origin of the 16.3-day periodicity (Sand et al., 2023).

A 2023 model-diagnosis study therefore argued that the 16.3-day periodicity itself is most naturally accommodated by an ultra-long rotation model, because the period is stable and the chromatic active window can be reproduced geometrically. However, the same paper found that an isolated ultra-long-period neutron star does not naturally explain the long-term RM evolution, and it proposed instead a hybrid model: a slowly rotating neutron star in a wide, eccentric massive-star binary, where the 16-day clock is rotational but the RM variations are produced by propagation through the companion’s magnetized wind. In that framework the preferred orbital period is $\alpha$8–3400 d, and the model predicts future periodic RM evolution on that much longer timescale (Lan et al., 2023).

Polarization-position-angle monitoring has added a further discriminator. A 2025 uGMRT study concluded that the PA varies with the periodicity of the source, is stable to within seven degrees on timescales less than four hours, and may vary between cycles even at the same activity phase. On that basis, the authors stated that the rotational model partially agrees with the observed PA variability, whereas all flavors of precessional models in which precession explains either the 16.34-day periodicity or the inter-cycle variability are robustly ruled out (Bethapudi et al., 10 Jul 2025). This is a model-dependent conclusion rather than a universally adopted consensus, but it substantially tightened the precession debate.

Earlier theoretical work had already made the periodicity problem explicit. One jet-precession study found that disk-driven jet precession requires an implausibly small viscosity parameter, $\alpha$9, to reproduce the 16.35-day period, whereas a tidal-force-driven precessing jet in a compact NS/BH+WD binary can produce periods of several days to hundreds of days (Chen et al., 2021). Another theoretical assessment argued more generally that the period derivative is a decisive discriminator among periodically modulated FRB models, because some classes predict measurable secular drift while others do not (Katz, 2020).

The broad synthesis remains provisional. The data strongly favor a picture in which FRB 20180916B has a stable $$L_{\mathrm{H\alpha}} \lesssim 8\times10^{36}\,\mathrm{erg\,s^{-1}}$$0-day visibility clock, a frequency-dependent activity window, narrowband burst spectra, and a local or host-galaxy Faraday screen whose magnetic configuration evolves largely independently of the dispersive column. This suggests that periodic burst visibility and long-term magneto-ionic evolution need not arise from the same physical component. Which dynamical framework best unifies those facts remains an active question.