Polarized light shows hot gas swirling around a galactic core

In 1918 US astronomer Heber Curtis peered at M87 from the Lick Observatory in California and reported “a curious straight ray … apparently connected with the nucleus by a thin line of matter.” Curtis was looking at a jet of plasma launched at relativistic speed from the galaxy’s center, although the theory to explain it came 30 years later. Radio observatories later linked the upstream end of the jet to a prominent radio source at M87’s center 55 million light-years from Earth in the Virgo galaxy cluster.

By the 1970s many theorists were convinced that such compact radio sources provided evidence that supermassive black holes occupy the centers of galaxies, including M87 and our own Milky Way, just as general relativity predicted. More specifically, Einstein’s field equations describe the geometry of empty spacetime around a so-called Kerr black hole, one that spins about a central axis and is characterized entirely by its mass and angular momentum.

Inspired by Roger Penrose’s theory of black hole formation, Roger Blandford and Roman Znajek suggested that jets emerge from galaxies as the outflow from an accretion disk—the region in which magnetized gas swirls around the event horizon.³ (See Physics Today, December 2020, page 14.) The spinning black hole acts like a magnetized conductor whose rotational energy is extracted and expelled in the form of jets.

Tracing the dynamics of those accretion and outflow processes, however, requires directly observing the immediate surroundings of a black hole. But the small-scale structure of a black hole’s event horizon, tens of millions of light-years away, has the same angular size as would an apple on the Moon as seen from Earth. “To a radio astronomer in the 1960s and 1970s, the idea that you could resolve at this scale was complete fantasy. Only the most visionary would dream about it,” says Blandford. But the development of very long baseline interferometry (VLBI), which combines observations from multiple radio antennas to create images of regions in space that are difficult to observe with a single antenna, led others to declare imaging a black hole as merely “a very hard thing to do,” he says.

VLBI depends on a network of radio antennas. For the EHT, those antennas are spread across four continents to form a virtual telescope with an Earth-sized aperture. Each antenna records signals emitted by the target source, and the signals’ arrival times are later cross-correlated for each pair of antennas. From those correlated signals, an image of the emission source can later be reconstructed.

The angular resolution of an array increases not only with the distance, or baseline, between antennas but also with the frequency of the observed radiation. At frequencies of 230 GHz, corresponding to a wavelength of 1.3 mm, Earth-sized baselines achieve the highest angular resolution in ground-based astronomy—sufficient to observe supermassive black holes in M87 and the Milky Way at the scale of the event horizon.

The EHT project, started in 2009, comprises a VLBI network of millimeter-wavelength radio telescopes, shown in figure 2. Those telescopes comprise arrays of individual radio dishes with baselines ranging from hundreds to thousands of kilometers.

Over five nights in April 2017, the EHT monitored radio emission from the accretion disk around M87. There, spacetime curvature causes electromagnetic radiation to bend and twist. Electrons tightly orbit the magnetic field lines; as a result, the emitted radio waves are polarized by the orientation of the original fields.⁴ Revisiting the 2017 data to establish measurements of the light’s polarization, the collaboration has now published images that show the pattern of magnetic fields near M87’s core, plotted in figure 1b. “It’s like looking at a medical x ray. All of a sudden you can see the underlying skeleton that supports everything,” says Harvard University’s Sheperd Doeleman, a collaborator on the EHT.

In one find, the total level of polarization measured was far below that expected given the fundamental processes involved. Although the radiation emitted by thermal electrons as they gyrate around magnetic field lines should be polarized, not all the light that travels to Earth is fully polarized. That’s because as the radiation makes its way through the plasma along the observer’s line of sight, another electromagnetic process—Faraday rotation—comes into play.

If all the magnetic field lines around M87 were neatly aligned, the radiation’s polarization would be strong and rotated by the same amount. But if the field lines were instead tangled together, whatever polarization the radiation originally had would be reduced. The EHT’s observation of up to 20% polarization in the brightest regions, then, is consistent with fields being tangled on small scales.

Another clue about the magnetic field’s structure came from the polarization’s orientation. Rather than a radial pattern, which would imply a field that followed the plasma and encircled the black hole, the polarization appeared with an azimuthal twist, seen in figure 1a. That distribution implies that electric fields are distributed in a spiral around the center and the magnetic field topology is poloidal—that is, along the direction of the jet.

With those observations, the EHT researchers could piece together the processes by which the black hole funnels material to its center. EHT team member Monika Moscibrodzka of Radboud University and her colleagues have simulated a library of templates that represent 120 models of accretion flows and jets, especially their emissivity properties. From the emissivity, the researchers calculated millions of light trajectories that originate at points on the accretion disk and travel through curved spacetime until arriving at telescopes on Earth.

Comparing a gallery of 72 000 simulated images to the new observations showed that the data were compatible with the magnetically arrested disk (MAD) model.⁴ In that model, magnetic flux accumulates and intensifies in the accreting gas. The field generates a pressure that’s large enough to reduce the accretion rate and create a bottleneck in the disk that pushes back against the inflow.⁵ 

From the MAD models, the EHT researchers estimated that M87’s 6-billion-solar-mass black hole swallows the surrounding gas at a rate of 1/1000th the mass of our sun per year. Still, that’s powerful enough to launch relativistic jets of charged particles that stretch for thousands of light-years, according to energy conservation mechanisms first proposed in the 1970s.

Observations by the Atacama Large Millimeter/Submillimeter Array (ALMA) in Chile, whose 66 antennas dominate the overall signal collection of the EHT, captured images of the jet in polarized light, shown in figure 3. Those observations allowed the researchers to infer polarization and magnetic field properties along the jet on scales of up to 6000 light-years.⁶ “Based on standard models, we expect the jet to be confined by very strong magnetic fields at its base, close to its launching point,” says Ciriaco Goddi, who works at Radboud University and with the EHT collaboration. The combined information from ALMA and the EHT establishes that link.

The next-generation EHT will produce movies of the jet launch region and the magnetic field structures that propel matter outwards from M87. Showing dynamically how magnetic fields launch jets will answer the next big question: How exactly do magnetic fields extract energy from a spinning black hole?

More telescopes are already being added to the EHT. The collaboration will observe at more wavelengths and will target other galaxies and black holes. Says Blandford, “This is the start of the story, not the end.”