Structures in the Interstellar Medium

Image: the original extreme scattering event. (Credit: Fiedler et al. 1987)

  • The interstellar medium is filled with turbulent, free electrons that cause variation to propagating radio signals. Common optical effects include dispersion (like a prism splitting visible light, the radio light splits and takes longer to arrive at our telescopes depending on the wavelength), refraction (bulk bending of light), scintillation (like the twinkling of starlight), scattering (diffuse deviations in the propagation path), and more.
  • Structures in the interstellar medium have been seen in various forms, including "extreme scattering events" which have typically been seen as a change in the brightness of the pulsar over long timescales. They have also been observed in variations of "scintillation parameters".
  • Precision timing of millisecond pulsars over years and decades allows us to build models of radio-pulse arrival times of incredible accuracy. These models account for the pulsar's spin period, its evolution, the dispersion of radio pulses in the interstellar medium, and many other effects. Developing precision timing models for many pulsars allows for surveying many lines of sight through the galaxy.
  • Advancing the modeling and correction of interstellar delays in our timing data is a prime focus of NANOGrav's Interstellar Medium Mitigation working group.

The NANOGrav 12.5-year Preliminary Data Set

  • NANOGrav's upcoming data release contains radio-pulse times of arrival (TOAs) and timing-models for 48 millisecond pulsars.
  • The observations span roughly 12.9 years, from July 30, 2005, to June 30, 2017. The pulsar with the longest baseline is J1744-1134, with 12.87 years of observations. The pulsar with the largest data volume is J1713+0747, with 40,000 TOAs.
  • Observations were carried out using the 100-m Robert C. Byrd Green Bank Telescope (GBT) of the Green Bank Observatory, and the 305-m William E. Gordon Telescope (Arecibo) of Arecibo Observatory.
  • In the sky map shown here, pulsar positions are marked by circles, with areas proportional to the number of TOAs in the dataset; the color scale indicates the timing baseline. The 34 pulsars with baselines greater than 3 years have solid red edges. We use only these 34 in our searches for a GW background.

Image: sky map of NANOGrav pulsars in the 11-yr data set.

Chromatic Timing Residuals

Image: Timing residuals for our three frequency bands with no frequency-dependent ("chromatic") delays removed except for a single fiducial dispersion measure delay. Both events are clearly seen as dips in the 820~MHz (top panels) data.

  • Using our 11-year data release, we fixed the spin, astrometric, and binary parameters for the pulsar, along with a fiducial dispersion measure delay. To start, no parameters were fit for and we did not include any time-variable dispersion-measure modeling. The event occured after the end of the 11-year data release time and we did not want the event to influence the timing model used to show its existence. The continued development of an efficient data reduction pipeline allowed for the quick identification of this event in our preliminary 12.5-year data set.
  • We see the two timing events most clearly in the lowest frequency band with which we observe PSR J1713+0747. The amplitude of propagation effects typically increase with decreasing radio frequency. While the coverage of the first event was sparse, we were able to constrain the rapid change in the average arrival time to within four days.
  • While the events are separated by roughly 7.5 years, we did not see any other dips in the 21-year timing of the pulsar, ruling out a possible periodic origin. The large amplitude, rapid decrease implies that it cannot be the pulsar moving through some underdensity of material local to the pulsar.
  • Even though the dips appear to have some "exponential" functional form, there is no reason to believe that the cause of the events did not result in richer structures in our timeseries.

Modeling of the Timing Delays

  • We chose to model the timing delays in two ways. The first was with the standard NANOGrav methodology of assuming a constant (frequency)−2 dependence over small time windows meant to model dispersion measure variations (see image). The second was by fitting a physical model that accounted for turbulent variations, the changing line of sight through the interstellar medium from pulsar-Earth motions, the solar wind, and more.
  • The second event occurs when we would have expected an increase in the dispersion measure from the line of sight passing through the solar wind close to the Sun yet we still see a large dip at that time. The timeseries shows many correlated structures that can be due to a variety of different causes.
  • The pulsar traces a path across the sky that samples different lines of sight through the interstellar medium. No unusual structures were identified in infrared or optical data of this sky position.
  • By considering the fact that no other such events have been seen in any other pulsar timeseries (at these amplitudes), we believe that the two events cannot be from independent structures or else we would expect many more to be seen in our data for other pulsars. Therefore we expect that the two events are linked to a single structure.

Image: Modeling of the arrival-time delays with the standard NANOGrav methodology. We also show the trajectory of the pulsar on the sky with the delay amplitudes overlaid.

The Lensing Model

Image: A schematic representation of the light ray paths refracting due to a Gaussian plasma lens (Credit: Clegg, Fey, & Lazio 1998).

  • Extreme scattering events are the result of "plasma lensing" in which the radio emission is refracted by some structure in the interstellar medium. Work has been done previously to understand lenses with a symmetric Gaussian structure (see image) though the results apply more generally.
  • Lensing can result from either an overdensity or an underdensity of electrons in the interstellar medium. The lensing causes additional non-dispersive delays that we must account for if we hope to properly model our data.
  • We saw marginal evidence for non-(frequency)−2 delays around the time of the second event. However, lensing can produce a wide variety of patterns in our timeseries, it is unclear at this time how to properly test for other dependencies in frequency, which will need to be the focus of future work for all of our pulsars.
  • While we do not claim the detection of a plasma lens, this is the only model that is consistent with our observations at this time.

Future Prospects for Event Detection

  • Studying these events as they start to happen or soon after they start to happen is of importance to understanding their origins and impacts. Targeted observations after an event has been identified can help increase cadence, integration time, and frequency coverage to gather as much information as possible on the events.
  • NANOGrav's Cyber-Infrastructure working group has developed automated transfer of data from the telescopes to our computing resources. The pipelines to reduce the data and produce usable arrival times are nearly automated. These advances in our computing capabilities will allow for near-real-time monitoring of these events.
  • We have developed software called Quicklook which allows for fast display of recently acquired pulsar observations. The plots provide a variety of metrics for diagnosing whether or not changes in timing due to the interstellar medium may be occuring.
  • A more rigorous analysis of our current data set may show more of these events in the timing data of other pulsars but at lower amplitude. The impact of these events on our gravitational wave sensitivity is an area of active research.

Image: An example of the Quicklook plotting output, which are currently being produced as part of our automated processing pipeline.


  • Members of the NANOGrav Collaboration: M. T. Lam, J. A. Ellis, G. Grillo, M. L. Jones, J. S. Hazboun, P. R. Brook, J. E. Turner, S. Chatterjee, J. M. Cordes, T. J. W. Lazio, M. E. DeCesar, Z. Arzoumanian, H. Blumer, H. T. Cromartie, P. B. Demorest, T. Dolch, R. D. Ferdman, E. C. Ferrara, E. Fonseca, N. Garver-Daniels, P. A. Gentile, V. Gupta, D. R. Lorimer, R. S. Lynch, D. R. Madison, M. A. McLaughlin, C. Ng, D. J. Nice, T. T. Pennucci, S. M. Ransom, R. Spiewak, I. H. Stairs, D. R. Stinebring, K. Stovall, J. K. Swiggum, S. J. Vigeland, W. W. Zhu
  • Contact: Dr. Michael T. Lam (corresponding author), Prof. Maura McLaughlin (NANOGrav chair).

The NANOGrav Collaboration at the 2017 Fall meeting in Lafayette College, PA