Extremely Variable Quasars

The "changing-look" quasars are likely a biased subset of the extremely variable quasar population.

Research Highlight: Demystifying Extreme Quasars

Quasars, the brilliantly luminous cores of distant galaxies, are powered by supermassive black holes actively devouring matter. One of their defining features is that they “flicker”—their brightness varies over time. For most quasars, this variation is relatively modest.

However, a small fraction exhibit truly dramatic changes. Some, known as Extreme Variability Quasars (EVQs), can change in brightness by more than a factor of 2.5 (> 1 magnitude). An even more enigmatic group, the “Changing-Look” Quasars (CLQs), undergo transformations so profound that they appear to change type entirely, with their characteristic broad emission lines—the “smoking gun” of an active quasar—seemingly vanishing or appearing out of nowhere.

Why extreme variability matters

This extreme behavior raises a fundamental question: Are these rare, dramatic objects (EVQs and CLQs) a truly distinct population, operating under a different set of physical rules? Or are they simply the most extreme examples of the same variability process we see in all quasars?

In our comprehensive study, we assembled the largest-ever sample of EVQs to connect their statistical, population-level properties with individual, multi-epoch behavior over time.

The Investigation: From a Global Census to Individual Changes

Our first step was to build a robust sample for study. By combining over 15 years of photometric data from the Sloan Digital Sky Survey (SDSS) and Pan-STARRS1, we constructed the largest catalog of its kind, identifying 14,012 EVQs (Ren et al. 2022).

A key challenge was that most of these quasars only had a single snapshot in time—one spectrum from SDSS. To overcome this, we developed a “state” classification. We compared the brightness of the quasar at the moment its spectrum was taken to its long-term average brightness. This allowed us to sort every spectrum into a state, from “Extremely Dim” to “Extremely Bright.” By “stacking” (averaging) thousands of spectra for each state, we could analyze the average properties of EVQs as they transitioned through their brightness cycles.

Key findings

  • A Familiar “Bluer-When-Brighter” Trend: Just like normal quasars, EVQs are “bluer” (emit more high-energy light) when they are bright and “redder” when they are dim. This was the first major clue that the underlying physical mechanism might be the same across all quasars.
  • The Intrinsic Baldwin Effect: We saw a clear Intrinsic Baldwin Effect (iBeff) for high-ionization broad lines like Mg II and C IV. This means the lines themselves became “weaker” (had a smaller equivalent width, or EW) as the quasar’s continuum brightness increased.
  • Intrinsically Stronger Lines: This was a crucial discovery. When we compared EVQs in their median state to a meticulously matched “control sample” of normal quasars, we found that EVQs have systematically stronger emission lines. This isn’t an illusion caused by their variability; it’s an intrinsic property. It suggests that the physics of EVQs—perhaps related to stronger disk turbulence or a harder ionizing spectrum—is quantitatively different, even if qualitatively similar.

The H$\beta$ puzzle and the CLQ connection

Stacked spectra initially suggested an anomaly: the broad H$\beta$ line did not show the expected Baldwin effect. Its strength remained almost constant across all brightness states.

To solve this, we had to zoom in. We shifted our focus to a subsample of 1,259 EVQs that had multiple SDSS spectra, allowing us to watch them change individually over time (Ren et al. 2024). This dynamic perspective was the key.

We discovered that the H$\beta$ anomaly was caused by host galaxy contamination. In the dim states, the light from the quasar’s much fainter host galaxy was “polluting” the spectrum, making the H$\beta$ line appear weaker than it really was. When we either isolated the most luminous quasars (where the host was drowned out) or applied a correction, the normal H$\beta$ iBeff reappeared. The puzzle was solved.

This insight clarifies the nature of CLQs. Our (host-corrected) EVQs show a normal iBeff for H$\beta$ (the line strengthens as the continuum dims), whereas literature CLQs show no iBeff after host corrected. The difference arises from selection bias: CLQs are defined by the apparent “disappearance” of broad H$\beta$ in the dim state—a qualitative criterion far more likely when the host galaxy is bright enough to hide the H$\beta$ line. By doing a rough host removal, we effectively mitigate the anomalous behavior of H$\beta$ line in CLQs, which is expected to be further improve if with better decomposition method. Consistently, those same CLQs show a normal Mg II iBeff where host contamination is relatively low, matching EVQs.

Overall conclusion

EVQs are part of the normal quasar population, representing the extreme tail of variability and characterized by intrinsically stronger emission lines. The “Changing-Look” phenomenon is not a distinct physical transformation; it is extreme continuum variability viewed through a biased definition sensitive to host contamination.

Resources