---
title: "How to read the CHIME paper on the 16-day FRB cycle"
description: "Read the CHIME/FRB discovery of 16.35-day periodic activity figure by figure, and sort its binary, magnetar and lensing explanations by evidence."
canonical_url: "https://sandboxphysics.com/stories/fast-radio-burst-echoes/reading-the-chime-r3-periodicity-paper/"
language: "en"
page_type: "story"
alternate_languages:
  zh-CN: "https://sandboxphysics.com/zh/stories/fast-radio-burst-echoes/reading-the-chime-r3-periodicity-paper/"
last_modified: "2026-11-14"
---

[Fast radio bursts: repeats, echoes and plasma lenses](https://sandboxphysics.com/stories/fast-radio-burst-echoes/)

# How to read the CHIME paper on the 16-day FRB cycle

Sandbox Physics · 2026-11-14

The paper measures one thing firmly: 38 bursts from FRB 20180916B, recorded between September 2018 and February 2020, all arrive inside a 5-day window that repeats every $16.35\pm0.15$ days. It does not find the cause. Orbital motion, absorption or lensing in a companion's wind, and precession are offered as possibilities, and a purely sporadic source is ruled out. Read the three main figures in order, then sort every later sentence into measurement, model or prediction.

This guide uses arXiv:2001.10275v5, the version we read, with its figure numbers. The paper was published in Nature in 2020 (volume 582, pages 351 to 355); we did not compare the journal text. [Periodic activity from a fast radio burst source](https://arxiv.org/abs/2001.10275v5) Before reading, it helps to know [how timing statistics separate source and path](https://sandboxphysics.com/stories/fast-radio-burst-echoes/repetition-statistics-source-or-path/).

## What the paper claims, in one paragraph

The CHIME/FRB Collaboration folded the barycentric, dedispersed arrival times of 38 bursts at trial periods from 1.5 to 100 days. A peak at 16.35 days has a chance probability of about $10^{-7}$ after accounting for the number of periods searched. All bursts fall in a 5-day phase window and half in a 0.6-day window. The detection rate inside $\pm2.7$ days of the activity peak is $0.9^{+0.5}_{-0.4}$ bursts per hour above 5.2 Jy ms; outside, the 95 percent upper limit is 0.07 per hour. The authors say the period could also be a higher-frequency alias.

## Figure 1: the periodogram and its controls

Panel a shows the reduced $\chi^2$ of the folded profile, with 5 phase bins, against trial period. Only bursts separated by at least a sidereal day count as independent, to allow for clustering. The tall peak is 16.35 days; arrows mark its subharmonics, and the 32.7-day peak is the next most prominent.

Panels b and c are the controls, and they are what make panel a believable. Panel b folds mock arrival times drawn from the telescope's daily exposure to the source. Panel c folds pulses from the Galactic pulsar B2319+60, seen by the same instrument and software. Neither shows a 16.35-day peak. In $10^6$ control sets of each type, no period reaches the significance of the real one.

Ask while reading: what would CHIME's once-a-day transit do to a period? The text answers that the daily exposure makes certain periods of hours to a day degenerate with 16.35 days. The significance of a periodicity is unaffected, but which period is the true one needs data taken at other times of day.

## Figure 2: when CHIME was looking

Panel a plots the daily exposure, the detections (with counts per active phase) and the shaded $\pm2.7$-day windows. Panel b shows the daily noise level measured from pulsars. Together they answer the obvious objection that the bursts only appear when the telescope happens to look. CHIME had substantial exposure and nominal sensitivity well outside the windows and saw nothing there. The figure also marks a European VLBI Network detection at 1.7 GHz near the leading edge of a window and a 17.6-hour Effelsberg non-detection at 1.4 GHz during a predicted peak, while CHIME detected two bursts at the same epoch.

## Figure 3: what changes with phase

Panels a and b plot fluence and dispersion measure against phase, folded at 16.35 days with MJD 58369.18 as phase zero; phase 0.5 is the mean of the folded arrivals. Panel c counts bursts per phase and panel d gives the exposure per phase bin. The text draws two conclusions. Fluence shows no trend with phase. Four bursts with baseband data, plus the EVN burst, give consistent dispersion measures, limiting changes to below $0.1\,\mathrm{pc\,cm^{-3}}$ over about 176 days; the noisier intensity-data values disfavour changes above 2 over the full span.

Read this figure with the lens chapters in mind. A plasma lens gives copies with different apparent dispersion measures, by up to the lens column. Stable dispersion among these five bursts limits any lens whose copies would differ by more than about 0.1; weaker lenses remain allowed.

## Extended Data: the bursts themselves

Extended Data Figs. 1 to 4 show the dynamic spectra. The main text notes that nearly all bursts show downward-drifting sub-bursts, at all phases. Keep that in mind for the lensing discussion below.

## Sorting the discussion

**Table 6.1** Claims in the paper by kind of evidence.

| Claim | Kind |
| --- | --- |
| 16.35-day periodicity, 5-day window, rate inside and outside it | Measured |
| No trend of fluence, width or bandwidth with phase; drifting sub-bursts at all phases | Measured |
| DM stable to $0.1\,\mathrm{pc\,cm^{-3}}$ over 176 days | Measured, from 5 bursts |
| Purely sporadic emission, such as isolated giant pulses, is excluded | Inference from the period |
| Orbital motion in an eccentric binary, wind or disk interaction, lower opacity at apastron | Proposed mechanisms |
| Lensing by companion wind clumps near periastron | Proposed, with a stated difficulty |
| Rotation of an isolated magnetar | Disfavoured: known magnetars spin in under 12 s |
| Precession of an isolated neutron star | Possible, with open physics |
| Phase-dependent DM or rotation measure, X-ray or gamma-ray emission, polarization angle changes | Predictions |

The lensing proposal cites Main and colleagues, who saw plasma lensing brighten the black widow pulsar PSR B1957+20 by factors up to 70 to 80 near eclipse. [Pulsar emission amplified and resolved by plasma lensing in an eclipsing binary](https://arxiv.org/abs/1805.09348v1) That is a measurement in a different system. The CHIME authors add that, if lensing were at work, the preference for downward-drifting frequency structure would be puzzling. Hessels and colleagues explain why: a changing lens geometry should make the drift change rate and sign, as it does in PSR B1957+20, where many lenses act. [FRB 121102 bursts show complex time-frequency structure, Sec. 4.3](https://arxiv.org/abs/1811.10748v1) Their sentence points to "Fig. 1", which in this version is the periodogram; the drifting spectra are in Extended Data Figs. 1 to 4. We read it as a reference to those spectra but have not checked the journal version.

Cordes and colleagues give the timescales to compare: a single moving lens produces a gain spike of about a day and a trough lasting months, with narrow, epoch-dependent spectral peaks. [Lensing of fast radio bursts by plasma structures in host galaxies](https://arxiv.org/abs/1703.06580v1) A lens could therefore shape when bursts cross a detection threshold, but a strictly repeating 16-day window would need a lens that returns to the line of sight every orbit, as in the binary pictures.

Shortly after the first version of this paper appeared, Lyutikov, Barkov and Giannios proposed that the cycle comes from absorption in an O or B star's wind around a neutron star. [FRB-periodicity: mild pulsars in tight O/B-star binaries](https://arxiv.org/abs/2002.01920v5) It was published in ApJ Letters in 2020. We read only its abstract and cite it as a proposed model.

## What later data say, and what stays open

Our tutorial folds 104 events from the second CHIME catalogue, extending to MJD 60126. This is our check, not a result of the paper. 91 percent fall inside the published window at 16.35 days. The phase uncertainty from the published $\pm0.15$ days grows to about one cycle by the last event, so these later bursts cannot be phased reliably with the 2020 ephemeris alone. A simple Rayleigh scan peaks at 16.33 days. The details are in [the method guide](https://sandboxphysics.com/stories/fast-radio-burst-echoes/fit-a-chime-burst-and-r3-repeats/).

Open questions the paper itself names:

- Whether 16.35 days is the true period or an alias, which data outside CHIME's daily windows can settle.
- How burst activity depends on frequency, given the 1.7 GHz detection at the leading edge of the window and the 1.4 GHz non-detection at a predicted peak.
- Which mechanism sets the window; phase-dependent DM or rotation measure would point to a companion's wind.

## Lab tasks

In [the plasma-lens echo Lab](https://sandboxphysics.com/labs/frb-plasma-lens-echoes/), open the delay chart and compare the lens-maximum line with the observed repeat gaps, many of them close to 16 days. Then choose the caustic preset and move the lens slowly. Note how the summed gain rises and falls, and ask how many such swings a single lens could give in one 16-day cycle.

## Try it in the Lab

[FRB echoes through a plasma lens](https://sandboxphysics.com/labs/frb-plasma-lens-echoes/): Use the delay chart to see where a lens echo would sit beside a 16-day cycle.

## Sources

1. [Periodic activity from a fast radio burst source](https://arxiv.org/abs/2001.10275v5)
2. [Pulsar emission amplified and resolved by plasma lensing in an eclipsing binary](https://arxiv.org/abs/1805.09348v1)
3. [FRB 121102 bursts show complex time-frequency structure, Sec. 4.3](https://arxiv.org/abs/1811.10748v1)
4. [Lensing of fast radio bursts by plasma structures in host galaxies](https://arxiv.org/abs/1703.06580v1)
5. [FRB-periodicity: mild pulsars in tight O/B-star binaries](https://arxiv.org/abs/2002.01920v5)

## Keep reading

Fast radio bursts: repeats, echoes and plasma lenses · Reading path

### From one burst to source-or-path tests

1. [Article **Is a repeating radio flash the source repeating, or its path?**](https://sandboxphysics.com/stories/fast-radio-burst-echoes/repeat-or-echo/)
2. [Article **How a clump of plasma delays and bends a radio burst**](https://sandboxphysics.com/stories/fast-radio-burst-echoes/plasma-delay-and-bending/)
3. [Article **Images, gains and delays of a Gaussian plasma lens**](https://sandboxphysics.com/stories/fast-radio-burst-echoes/gaussian-plasma-lens-images/)
4. [Article **Telling source repetition from path effects with timing statistics**](https://sandboxphysics.com/stories/fast-radio-burst-echoes/repetition-statistics-source-or-path/)
5. [Article **Search one CHIME burst for a lens echo and test R3's repeats**](https://sandboxphysics.com/stories/fast-radio-burst-echoes/fit-a-chime-burst-and-r3-repeats/)
6. [You are here **How to read the CHIME paper on the 16-day FRB cycle**](https://sandboxphysics.com/stories/fast-radio-burst-echoes/reading-the-chime-r3-periodicity-paper/)
